Location and tracking of mobile devices: Uberveillance stalks the streets

Abstract

During the last decade, location-tracking and monitoring applications have proliferated, in mobile cellular and wireless data networks, and through self-reporting by applications running in smartphones that are equipped with onboard global positioning system (GPS) chipsets. It is now possible to locate a smartphone user's location not merely to a cell, but to a small area within it. Innovators have been quick to capitalise on these location-based technologies for commercial purposes, and have gained access to a great deal of sensitive personal data in the process. In addition, law enforcement utilises these technologies, can do so inexpensively and hence can track many more people. Moreover, these agencies seek the power to conduct tracking covertly, and without a judicial warrant. This article investigates the dimensions of the problem of people-tracking through the devices that they carry. Location surveillance has very serious negative implications for individuals, yet there are very limited safeguards. It is incumbent on legislatures to address these problems, through both domestic laws and multilateral processes.

1. Introduction

Personal electronic devices travel with people, are worn by them, and are, or soon will be, inside them. Those devices are increasingly capable of being located, and, by recording the succession of locations, tracked. This creates a variety of opportunities for the people concerned. It also gives rise to a wide range of opportunities for organisations, at least some of which are detrimental to the person's interests.

Commonly, the focus of discussion of this topic falls on mobile phones and tablets. It is intrinsic to the network technologies on which those devices depend that the network operator has at least some knowledge of the location of each handset. In addition, many such devices have onboard global positioning system (GPS) chipsets, and self-report their coordinates to service-providers. The scope of this paper encompasses those already well-known forms of location and tracking, but it extends beyond them.

The paper begins by outlining the various technologies that enable location and tracking, and identifies those technologies' key attributes. The many forms of surveillance are then reviewed, in order to establish a framework within which applications of location and tracking can be characterised. Applications are described, and their implications summarised. Controls are considered, whereby potential harm to the interests of individuals can be prevented or mitigated.

2. Relevant technologies

The technologies considered here involve a device that has the following characteristics:

• it is conveniently portable by a human, and

• it emits signals that:

• enable some other device to compute the location of the device (and hence of the person), and

• are sufficiently distinctive that the device is reliably identifiable at least among those in the vicinity, and hence the device's (and hence the person's) successive locations can be detected, and combined into a trail

The primary form-factors for mobile devices are currently clam-shape (portable PCs), thin rectangles suitable for the hand (mobile phones), and flat forms (tablets). Many other form-factors are also relevant, however. Anklets imposed on dangerous prisoners, and even as conditions of bail, carry RFID tags. Chips are carried in cards of various sizes, particularly the size of credit-cards, and used for tickets for public transport and entertainment venues, aircraft boarding-passes, toll-road payments and in some countries to carry electronic cash. Chips may conduct transactions with other devices by contact-based means, or contactless, using radio-frequency identification (RFID) or its shorter-range version near-field communication (NFC) technologies. These capabilities are in credit and debit cards in many countries. Transactions may occur with the cardholder's knowledge, with their express consent, and with an authentication step to achieve confidence that the person using the card is authorised to do so. In a variety of circumstances, however, some and even all of those safeguards are dispensed with. The electronic versions of passports that are commonly now being issued carry such a chip, and have an autonomous communications capability. The widespread issue of cards with capabilities uncontrolled by, and in many cases unknown to, the cardholder, is causing consternation among segments of the population that have become aware of the schemes.

Such chips can be readily carried in other forms, including jewellery such as finger-rings, and belt-buckles. Endo-prostheses such as replacement hips and knees and heart pacemakers can readily carry chips. A few people have voluntarily embedded chips directly into their bodies for such purposes as automated entry to premises (Michael and Michael, 2009).

In order to locate and track such devices, any sufficiently distinctive signals may in principle suffice. See Raper et al. (2007a) and Mautz (2011). In practice, the signals involved are commonly those transmitted by a device in order to take advantage of wireless telecommunications networks. The scope of the relevant technologies therefore also encompasses the signals, devices that detect the signals, and the networks over which the data that the signals contain are transmitted.

In wireless networks, it is generally the case that the base-station or router needs to be aware of the identities of devices that are currently within the cell. A key reason for this is to conserve limited transmission capacity by sending messages only when the targeted device is known to be in the cell. This applies to all of:

• cellular mobile originally designed for voice telephony and extended to data (in particular those using the ‘3G’ standards GSM/GPRS, CDMA2000 and UMTS/HSPA and the ‘4G’ standard LTE)

• wireless local area networks (WLANs, commonly Wifi/IEEE 802.11x – RE, 2010a)

• wireless wide area networks (WWANs, commonly WiMAX/IEEE 802.16x – RE, 2010b).

Devices in such networks are uniquely identified by various means (Clarke and Wigan, 2011). In cellular networks, there is generally a clear distinction between the entity (the handset) and the identity it is adopting at any given time (which is determined by the module inserted in it). Depending on the particular standards used, what is commonly referred to as ‘the SIM-card’ is an R-UIM, a CSIM or a USIM. These modules store an International Mobile Subscriber Identity (IMSI), which constitutes the handset's identifier. Among other things, this enables network operators to determine whether or not to provide service, and what tariff to apply to the traffic. However, cellular network protocols may also involve transmission of a code that distinguishes the handset itself, within which the module is currently inserted. A useful generic term for this is the device ‘entifier’ (Clarke, 2009b). Under the various standards, it may be referred to as an International Mobile Equipment Identity (IMEI), ESN, or MEID.

Vendor-specific solutions also may provide additional functionality to a handset unbeknown to the end-user. For example, every mobile device manufactured by Apple has a 40-character Unique Device Identifier (UDID). This enables Apple to track its users. Not only Apple itself, but also marketers, were able to use the UDID to track devices. It has also been alleged that data emanating from these devices is routinely accessible to law enforcement agencies. Since late 2012, Apple has prevented marketers from using the UDID, but has added an Identifier for Advertisers (IFA or IDFA). This is temporary, and it can be blocked; but it is by default open for tracking, and turning it off is difficult, and is likely to result in reduced services (Edwards, 2012). In short, Apple devices are specifically designed to enable tracking of consumers by Apple, by any government agency that has authority to gain access to the data, and by all consumer-marketing corporations, although in the last case with a low-grade option available to the user to suppress tracking.

In Wifi and WiMAX networks, the device entifier may be a processor-id or more commonly a network interface card identifier (NIC Id). In various circumstances, other device-identifiers may be used, such as a phone number, or an IP-address may be used as a proxy. In addition, the human using the device may be directly identified, e.g. by means of a user-account name.

A WWAN cell may cover a large area, indicatively of a 50 km radius. Telephony cells may have a radius as large as 2–3 km or as little as a hundred metres. WLANs using Wifi technologies have a cell-size of less than 1 ha, indicatively 50–100 m radius, but in practice often constrained by environmental factors to only 10–30 m.

The base-station or router knows the identities of devices that are within its cell, because this is a technically necessary feature of the cell's operation. Mobile devices auto-report their presence 10 times per second. Meanwhile, the locations of base-stations for cellular services are known with considerable accuracy by the telecommunications providers. And, in the case of most private Wifi services, the location of the router is mapped to c. 30–100 m accuracy by services such as Skyhook and Google Locations, which perform what have been dubbed ‘war drives’ in order to maintain their databases – in Google's case in probable violation of the telecommunications interception and/or privacy laws of at least a dozen countries (EPIC, 2012).

Knowing that a device is within a particular mobile phone, WiMAX or Wifi cell provides only a rough indication of location. In order to generate a more precise estimate, within a cell, several techniques are used (McGuire et al., 2005). These include the following (adapted from Clarke and Wigan, 2011; see also Figueiras and Frattasi, 2010):

• directional analysis. A single base-station may comprise multiple receivers at known locations and pointed in known directions, enabling the handset's location within the cell to be reduced to a sector within the cell, and possibly a narrow one, although without information about the distance along the sector;

• triangulation. This involves multiple base-stations serving a single cell, at known locations some distance apart, and each with directional analysis capabilities. Particularly with three or more stations, this enables an inference that the device's location is within a small area at the intersection of the multiple directional plots;

• signal analysis. This involves analysis of the characteristics of the signals exchanged between the handset and base-station, in order to infer the distance between them. Relevant signal characteristics include the apparent response-delay (Time Difference of Arrival – TDOA, also referred to as multilateration), and strength (Received Signal Strength Indicator – RSSI), perhaps supplemented by direction (Angle Of Arrival – AOA).

The precision and reliability of these techniques varies greatly, depending on the circumstances prevailing at the time. The variability and unpredictability result in many mutually inconsistent statements by suppliers, in the general media, and even in the technical literature.

Techniques for cellular networks generally provide reasonably reliable estimates of location to within an indicative 50–100 m in urban areas and some hundreds of metres elsewhere. Worse performance has been reported in some field-tests, however. For example, Dahunsi and Dwolatzky (2012) found the accuracy of GSM location in Johannesburg to be in the range 200–1400 m, and highly variable, with “a huge difference between the predicted and provided accuracies by mobile location providers”.

The website of the Skyhook Wifi-router positioning service claims 10-m accuracy, 1-s time-to-first-fix and 99.8% reliability (SHW, 2012). On the other hand, tests have resulted in far lower accuracy measures, including an average positional error of 63 m in Sydney (Gallagher et al., 2009) and “median values for positional accuracy in [Las Vegas, Miami and San Diego, which] ranged from 43 to 92 metres… [and] the replicability… was relatively poor” (Zandbergen, 2012, p. 35). Nonetheless, a recent research article suggested the feasibility of “uncooperatively and covertly detecting people ‘through the wall’ [by means of their WiFi transmissions]” (Chetty et al., 2012).

Another way in which a device's location may become known to other devices is through self-reporting of the device's position, most commonly by means of an inbuilt Global Positioning System (GPS) chipset. This provides coordinates and altitude based on broadcast signals received from a network of satellites. In any particular instance, the user of the device may or may not be aware that location is being disclosed.

Despite widespread enthusiasm and a moderate level of use, GPS is subject to a number of important limitations. The signals are subject to interference from atmospheric conditions, buildings and trees, and the time to achieve a fix on enough satellites and deliver a location measure may be long. This results in variability in its practical usefulness in different circumstances, and in its accuracy and reliability. Civil-use GPS coordinates are claimed to provide accuracy within a theoretical 7.8 m at a 95% confidence level (USGov, 2012), but various reports suggest 15 m, or 20 m, or 30 m, but sometimes 100 m. It may be affected by radio interference and jamming. The original and still-dominant GPS service operated by the US Government was subject to intentional degradation in the US's national interests. This ‘Selective Availability’ feature still exists, although subject to a decade-long policy not to use it; and future generations of GPS satellites may no longer support it.

Hybrid schemes exist that use two or more sources in order to generate more accurate location-estimates, or to generate estimates more quickly. In particular, Assisted GPS (A-GPS) utilises data from terrestrial servers accessed over cellular networks in order to more efficiently process satellite-derived data (e.g. RE, 2012).

Further categories of location and tracking technologies emerge from time to time. A current example uses means described by the present authors as ‘mobile device signatures’ (MDS). A device may monitor the signals emanating from a user's mobile device, without being part of the network that the user's device is communicating with. The eavesdropping device may detect particular signal characteristics that distinguish the user's mobile device from others in the vicinity. In addition, it may apply any of the various techniques mentioned above, in order to locate the device. If the signal characteristics are persistent, the eavesdropping device can track the user's mobile device, and hence the person carrying it. No formal literature on MDS has yet been located. The supplier's brief description is at PI (2010b).

The various technologies described in this section are capable of being applied to many purposes. The focus in this paper is on their application to surveillance.

3. Surveillance

The term surveillance refers to the systematic investigation or monitoring of the actions or communications of one or more persons (Clarke, 2009c). Until recent times, surveillance was visual, and depended on physical proximity of an observer to the observed. The volume of surveillance conducted was kept in check by the costs involved. Surveillance aids and enhancements emerged, such as binoculars and, later, directional microphones. During the 19th century, the post was intercepted, and telephones were tapped. During the 20th century, cameras enabled transmission of image, video and sound to remote locations, and recording for future use (e.g. Parenti, 2003).

With the surge in stored personal data that accompanied the application of computing to administration in the 1970s and 1980s, dataveillance emerged (Clarke, 1988). Monitoring people through their digital personae rather than through physical observation of their behaviour is much more economical, and hence many more people can be subjected to it (Clarke, 1994). The dataveillance epidemic made it more important than ever to clearly distinguish between personal surveillance – of an identified person who has previously come to attention – and mass surveillance – of many people, not necessarily previously identified, about some or all of whom suspicion could be generated.

Location data is of a very particular nature, and hence it has become necessary to distinguish location surveillance as a sub-set of the general category of dataveillance. There are several categories of location surveillance with different characteristics (Clarke and Wigan, 2011):

• capture of an individual's location at a point in time. Depending on the context, this may support inferences being drawn about an individual's behaviour, purpose, intention and associates

• real-time monitoring of a succession of locations and hence of the person's direction of movement. This is far richer data, and supports much more confident inferences being drawn about an individual's behaviour, purpose, intention and associates

• predictive tracking, by extrapolation from the person's direction of movement, enabling inferences to be drawn about near-future behaviour, purpose, intention and associates

• retrospective tracking, on the basis of the data trail of the person's movements, enabling reconstruction of a person's behaviour, purpose, intention and associates at previous times

Information arising at different times, and from different forms of surveillance, can be combined, in order to offer a more complete picture of a person's activities, and enable yet more inferences to be drawn, and suspicions generated. This is the primary sense in which the term ‘überveillance’ is applied: “Überveillance has to do with the fundamental who (ID), where (location), and when (time) questions in an attempt to derive why (motivation), what (result), and even how (method/plan/thought). Überveillance can be a predictive mechanism for a person's expected behaviour, traits, likes, or dislikes; or it can be based on historical fact; or it can be something in between… Überveillance is more than closed circuit television feeds, or cross-agency databases linked to national identity cards, or biometrics and ePassports used for international travel. Überveillance is the sum total of all these types of surveillance and the deliberate integration of an individual's personal data for the continuous tracking and monitoring of identity and location in real time” (Michael and Michael, 2010. See also Michael and Michael, 2007Michael et al., 20082010Clarke, 2010).

A comprehensive model of surveillance includes consideration of geographical scope, and of temporal scope. Such a model assists the analyst in answering key questions about surveillance: of what? for whom? by whom? why? how? where? and when? (Clarke, 2009c). Distinctions are also needed based on the extent to which the subject has knowledge of surveillance activities. It may be overt or covert. If covert, it may be merely unnotified, or alternatively express measures may be undertaken in order to obfuscate, and achieve secrecy. A further element is the notion of ‘sousveillance’, whereby the tools of surveillance are applied, by those who are commonly watched, against those who are commonly the watchers (Mann et al., 2003).

These notions are applied in the following sections in order to establish the extent to which location and tracking of mobile devices is changing the game of surveillance, and to demonstrate that location surveillance is intruding more deeply into personal freedoms than previous forms of surveillance.

4. Applications

This section presents a typology of applications of mobile device location, as a means of narrowing down to the kinds of uses that have particularly serious privacy implications. These are commonly referred to as location-based services (LBS). One category of applications provide information services that are for the benefit of the mobile device's user, such as navigation aids, and search and discovery tools for the locations variously of particular, identified organisations, and of organisations that sell particular goods and services. Users of LBS of these kinds can be reasonably assumed to be aware that they are disclosing their location. Depending on the design, the disclosures may also be limited to specific service-providers and specific purposes, and the transmissions may be secured.

Another, very different category of application is use by law enforcement agencies (LEAs). The US E-911 mandate of 1999 was nominally a public safety measure, to enable people needing emergency assistance to be quickly and efficiently located. In practice, the facility also delivered LEAs means for locating and tracking people of interest, through their mobile devices. Personal surveillance may be justified by reasonable grounds for suspicion that the subject is involved in serious crime, and may be specifically authorised by judicial warrant. Many countries have always been very loose in their control over LEAs, however, and many others have drastically weakened their controls since 2001. Hence, in any given jurisdiction and context, each and all of the controls may be lacking.

Yet worse, LEAs use mobile location and tracking for mass surveillance, without any specific grounds for suspicion about any of the many people caught up in what is essentially a dragnet-fishing operation (e.g. Mery, 2009). Examples might include monitoring the area adjacent to a meeting-venue watching out for a blacklist of device-identifiers known to have been associated with activists in the past, or collecting device-identifiers for use on future occasions. In addition to netting the kinds of individuals who are of legitimate interest, the ‘by-catch’ inevitably includes threatened species. There are already extraordinarily wide-ranging (and to a considerable extent uncontrolled) data retention requirements in many countries.

Of further concern is the use of Automated Number Plate Recognition (ANPR) for mass surveillance purposes. This has been out of control in the UK since 2006, and has been proposed or attempted in various other countries as well (Clarke, 2009a). Traffic surveillance is expressly used not only for retrospective analysis of the movements of individuals of interest to LEAs, but also as a means of generating suspicions about other people (Lewis, 2008).

Beyond LEAs, many government agencies perform social control functions, and may be tempted to conduct location and tracking surveillance. Examples would include benefits-paying organisations tracking the movements of benefits-recipients about whom suspicions have arisen. It is not too far-fetched to anticipate zealous public servants concerned about fraud control imposing location surveillance on all recipients of some particularly valuable benefit, or as a security precaution on every person visiting a sensitive area (e.g. a prison, a power plant, a national park).

Various forms of social control are also exercised by private sector organisations. Some of these organisations, such as placement services for the unemployed, may be performing outsourced public sector functions. Others, such as workers' compensation providers, may be seeking to control personal insurance claimants, and similarly car-hire companies and insurance providers may wish to monitor motor vehicles' distance driven and roads used (Economist, 2012Michael et al., 2006b).

A further privacy-invasive practice that is already common is the acquisition of location and tracking data by marketing corporations, as a by-product of the provision of location-based services, but with the data then applied to further purposes other than that for which it was intended. Some uses rely on statistical analysis of large holdings (‘data mining’). Many uses are, on the other hand, very specific to the individual, and are for such purposes as direct or indirect targeting of advertisements and the sale of goods and services. Some of these applications combine location data with data from other sources, such as consumer profiling agencies, in order to build up such a substantial digital persona that the individual's behaviour is readily influenced. This takes the activity into the realms of überveillance.

All such services raise serious privacy concerns, because the data is intensive and sensitive, and attractive to organisations. Companies may gain rights in relation to the data through market power, or by trickery – such as exploitation of a self-granted right to change the Terms of Service (Clarke, 2011). Once captured, the data may be re-purposed by any organisation that gains access to it, because the value is high enough that they may judge the trivial penalties that generally apply to breaches of privacy laws to be well worth the risk.

A recently-emerged, privacy-invasive practice is the application of the mobile device signature (MDS) form of tracking, in such locations as supermarkets. This is claimed by its providers to offer deep observational insights into the behaviour of customers, including dwell times in front of displays, possibly linked with the purchaser's behaviour. This raises concerns a little different from other categories of location and tracking technologies, and is accordingly considered in greater depth in the following section.

It is noteworthy that an early review identified a wide range of LBS, which the authors classified into mobile guides, transport, gaming, assistive technology and location-based health (Raper et al., 2007b). Yet that work completely failed to notice that a vast array of applications were emergent in surveillance, law enforcement and national security, despite the existence of relevant literature from at least 1999 onwards (Clarke, 2001Michael and Masters, 2006).

5. Implications

The previous sections have introduced many examples of risks to citizens and consumers arising from location surveillance. This section presents an analysis of the categories and of the degree of seriousness with which they should be viewed. The first topic addressed is the privacy of personal location data. Other dimensions of privacy are then considered, and then the specific case of MDS is examined. The treatment here is complementary to earlier articles that have looked more generally at particular applications such as location-based mobile advertising, e.g. Cleff (20072010) and King and Jessen (2010). See also Art. 29 (2011).

5.1. Locational privacy

Knowing where someone has been, knowing what they are doing right now, and being able to predict where they might go next is a powerful tool for social control and for chilling behaviour (Abbas, 2011). Humans do not move around in a random manner (Song et al., 2010).

One interpretation of ‘locational privacy’ is that it “is the ability of an individual to move in public space with the expectation that under normal circumstances their location will not be systematically and secretly recorded for later use” (Blumberg and Eckersley, 2009). A more concise definition is “the ability to control the extent to which personal location information is… [accessible and] used by others” (van Loenen et al., 2009). Hence ‘tracking privacy’ is the interest an individual has in controlling information about their sequence of locations.

Location surveillance is deeply intrusive into data privacy, because it is very rich, and enables a great many inferences to be drawn (Clarke, 2001Dobson and Fisher, 2003Michael et al., 2006aClarke and Wigan, 2011). As demonstrated by Raper et al. (2007a, p. 32–3), most of the technical literature that considers privacy is merely concerned about it as an impediment to deployment and adoption, and how to overcome the barrier rather than how to solve the problem. Few authors adopt a positive approach to privacy-protective location technologies. The same authors' review of applications (Raper et al., 2007b) includes a single mention of privacy, and that is in relation to just one of the scores of sub-categories of application that they catalogue.

Most service-providers are cavalier in their handling of personal data, and extravagant in their claims. For example, Skyhook claims that it “respects the privacy of all users, customers, employees and partners”; but, significantly, it makes no mention of the privacy of the people whose locations, through the locations of their Wifi routers, it collects and stores (Skyhook, 2012).

Consent is critical in such LBS as personal location chronicle systems, people-followers and footpath route-tracker systems that systematically collect personal location information from a device they are carrying (Collier, 2011c). The data handled by such applications is highly sensitive because it can be used to conduct behavioural profiling of individuals in particular settings. The sensitivity exists even if the individuals remain ‘nameless’, i.e. if each identifier is a temporary or pseudo-identifier and is not linked to other records. Service-providers, and any other organisations that gain access to the data, achieve the capacity to make judgements on individuals based on their choices of, for example, which retail stores they walk into and which they do not. For example, if a subscriber visits a particular religious bookstore within a shopping mall on a weekly basis, the assumption can be reasonably made that they are in some way affiliated to that religion (Samuel, 2008).

It is frequently asserted that individuals cannot have a reasonable expectation of privacy in a public space (Otterberg, 2005). Contrary to those assertions, however, privacy expectations always have existed in public places, and continue to exist (VLRC, 2010). Tracking the movements of people as they go about their business is a breach of a fundamental expectation that people will be ‘let alone’. In policing, for example, in most democratic countries, it is against the law to covertly track an individual or their vehicle without specific, prior approval in the form of a warrant. This principle has, however, been compromised in many countries since 2001. Warrantless tracking using a mobile device generally results in the evidence, which has been obtained without the proper authority, being inadmissible in a court of law (Samuel, 2008). Some law enforcement agencies have argued for the abolition of the warrant process because the bureaucracy involved may mean that the suspect cannot be prosecuted for a crime they have likely committed (Ganz, 2005). These issues are not new; but far from eliminating a warrant process, the appropriate response is to invest the energy in streamlining this process (Bronitt, 2010).

Privacy risks arise not only from locational data of high integrity, but also from data that is or becomes associated with a person and that is inaccurate, misleading, or wrongly attributed to that individual. High levels of inaccuracy and unreliability were noted above in respect of all forms of location and tracking technologies. In the case of MDS services, claims have been made of 1–2 m locational accuracy. This has yet to be supported by experimental test cases however, and hence there is uncertainty about the reliability of inferences that the service-provider or the shop owner draw. If the data is the subject of a warrant or subpoena, the data's inaccuracy could result in false accusations and even a miscarriage of justice, with the ‘wrong person’ finding themselves in the ‘right place’ at the ‘right time’.

5.2. Privacy more broadly

Privacy has multiple dimensions. One analysis, in Clarke (2006a), identifies four distinct aspects. Privacy of Personal Data, variously also ‘data privacy’ and ‘information privacy’, is the most widely discussed dimension of the four. Individuals claim that data about themselves should not be automatically available to other individuals and organisations, and that, even where data is possessed by another party, the individual must be able to exercise a substantial degree of control over that data and its use. The last five decades have seen the application of information technologies to a vast array of abuses of data privacy. The degree of privacy intrusiveness is a function of both the intensity and the richness of the data. Where multiple sources are combined, the impact is particularly likely to chill behaviour. An example is the correlation of video-feeds with mobile device tracking. The previous sub-section addressed that dimension.

Privacy of the Person, or ‘bodily privacy’, extends from freedom from torture and right to medical treatment, via compulsory immunisation and imposed treatments, to compulsory provision of samples of body fluids and body tissue, and obligations to submit to biometric measurement. Locational surveillance gives rise to concerns about personal safety. Physical privacy is directly threatened where a person who wishes to inflict harm is able to infer the present or near-future location of their target. Dramatic examples include assassins, kidnappers, ‘standover merchants’ and extortionists. But even people who are neither celebrities nor notorieties are subject to stalking and harassment (Fusco et al., 2012).

Privacy of Personal Communications is concerned with the need of individuals for freedom to communicate among themselves, without routine monitoring of their communications by other persons or organisations. Issues include ‘mail covers’, the use of directional microphones, ‘bugs’ and telephonic interception, with or without recording apparatus, and third-party access to email-messages. Locational surveillance thereby creates new threats to communications privacy. For example, the equivalent of ‘call records’ can be generated by combining the locations of two device-identifiers in order to infer that a face-to-face conversation occurred.

Privacy of Personal Behaviour encompasses ‘media privacy’, but particular concern arises in relation to sensitive matters such as sexual preferences and habits, political activities and religious practices. Some privacy analyses, particularly in Europe, extend this discussion to personal autonomy, liberty and the right of self-determination (e.g. King and Jessen, 2010). The notion of ‘private space’ is vital to economic and social aspects of behaviour, is relevant in ‘private places’ such as the home and toilet cubicles, but is also relevant and important in ‘public places’, where systematic observation and the recording of images and sounds are far more intrusive than casual observation by the few people in the vicinity.

Locational surveillance gives rise to rich sets of data about individuals' activities. The knowledge, or even suspicion, that such surveillance is undertaken, chills their behaviour. The chilling factor is vital in the case of political behaviour (Clarke, 2008). It is also of consequence in economic behaviour, because the inventors and innovators on whom new developments depend are commonly ‘different-thinkers’ and even ‘deviants’, who are liable to come to come to attention in mass surveillance dragnets, with the tendency to chill their behaviour, their interactions and their creativity.

Surveillance that generates accurate data is one form of threat. Surveillance that generates inaccurate data, or wrongly associates data with a particular person, is dangerous as well. Many inferences that arise from inaccurate data will be wrong, of course, but that won't prevent those inferences being drawn, resulting in unjustified behavioural privacy invasiveness, including unjustified association with people who are, perhaps for perfectly good reasons, themselves under suspicion.

In short, all dimensions of privacy are seriously affected by location surveillance. For deeper treatments of the topic, see Michael et al. (2006b) and Clarke and Wigan (2011).

5.3. Locational privacy and MDS

The recent innovation of tracking by means of mobile device signatures (MDS) gives rise to some issues additional to, or different from, mainstream device location technologies. This section accordingly considers this particular technique's implications in greater depth. Limited reliable information is currently available, and the analysis is of necessity based on supplier-published sources (PI, 2010a2010b) and media reports (Collier, 2011a,b,c).

Path Intelligence (PI) markets an MDS service to shopping mall-owners, to enable them to better value their floor space in terms of rental revenues, and to identify points of on-foot traffic congestion to on-sell physical advertising and marketing floor space (PI, 2010a). The company claims to detect each phone (and hence person) that enters a zone, and to capture data, including:

• how long each device and person stay, including dwell times in front of shop windows;

• repeat visits by shoppers in varying frequency durations; and

• typical route and circuit paths taken by shoppers as they go from shop to shop during a given shopping experience.

For malls, PI is able to denote such things as whether or not shoppers who shop at one establishment will also shop at another in the same mall, and whether or not people will go out of their way to visit a particular retail outlet independent of its location. For retailers, PI says it is able to provide information on conversion rates by department or even product line, and even which areas of the store might require more attention by staff during specific times of the day or week (PI, 2012).

PI says that it uses “complex algorithms” to denote the geographic position of a mobile phone, using strategically located “proprietary equipment” in a campus setting (PI, 2010a). The company states that it is conducting “data-driven analysis”, but is not collecting, or at least that it is not disclosing, any personal information such as a name, mobile telephone number or contents of a short message service (SMS). It states that it only ever provides aggregated data at varying zone levels to the shopping mall-owners. This is presumably justified on the basis that, using MDS techniques, direct identifiers are unlikely to be available, and a pseudo-identifier needs to be assigned. There is no explicit definition of what constitutes a zone. It is clear, however, that minimally-aggregated data at the highest geographic resolution is available for purchase, and at a higher price than more highly-aggregated data.

Shoppers have no relationship with the company, and it appears unlikely that they would even be aware that data about them is being collected and used. The only disclosure appears to be that “at each of our installations our equipment is clearly visible and labelled with our logo and website address” (PI, 2010a), but this is unlikely to be visible to many people, and in any case would not inform anyone who saw it.

In short, the company is generating revenue by monitoring signals from the mobile devices of people who visit a shopping mall for the purchase of goods and services. The data collection is performed without the knowledge of the person concerned (Renegar et al., 2008). The company is covertly collecting personal data and exploiting it for profit. There is no incentive or value proposition for the individual whose mobile is being tracked. No clear statement is provided about collection, storage, retention, use and disclosure of the data (Arnold, 2008). Even if privacy were not a human right, this would demand statutory intervention on the public policy grounds of commercial unfairness. The company asserts that “our privacy approach has been reviewed by the [US Federal Trade Commission] FTC, which determined that they are comfortable with our practices” (PI, 2010a). It makes no claims of such ‘approval’ anywhere else in the world.

The service could be extended beyond a mall and the individual stores within it, to for example, associated walkways and parking areas, and surrounding areas such as government offices, entertainment zones and shopping-strips. Applications can also be readily envisaged on hospital and university campuses, and in airports and other transport hubs. From prior research, this is likely to expose the individual's place of employment, and even their residence (Michael et al., 2006a,b). Even if only aggregated data is sold to businesses, the individual records remain available to at least the service-provider.

The scope exists to combine this form of locational surveillance with video-surveillance such as in-store CCTV, and indeed this is claimed to be already a feature of the company's offering to retail stores. To the extent that a commonly-used identifier can be established (e.g. through association with the person's payment or loyalty card at a point-of-sale), the full battery of local and externally acquired customer transaction histories and consolidated ‘public records’ data can be linked to in-store behaviour (Michael and Michael, 2007). Longstanding visual surveillance is intersecting with well-established data surveillance, and being augmented by locational surveillance, giving breath to dataveillance, or what is now being referred to by some as ‘smart surveillance’ (Wright et al., 2010IBM, 2011).

Surreptitious collection of personal data is (with exemptions and exceptions) largely against the law, even when undertaken by law enforcement personnel. The MDS mechanism also flies in the face of telephonic interception laws. How, then, can it be in any way acceptable for a form of warrantless tracking to be undertaken by or on behalf of corporations or mainstream government agencies, of shoppers in a mall, or travellers in an airport, or commuters in a transport hub? Why should a service-provider have the right to do what a law enforcement agency cannot normally do?

6. Controls

The tenor of the discussion to date has been that location surveillance harbours enormous threats to location privacy, but also to personal safety, the freedom to communicate, freedom of movement, and freedom of behaviour. This section examines the extent to which protections exist, firstly in the form of natural or intrinsic controls, and secondly in the form of legal provisions. The existing safeguards are found to be seriously inadequate, and it is therefore necessary to also examine the prospects for major enhancements to law, in order to achieve essential protections.

6.1. Intrinsic controls

A variety of forms of safeguard exist against harmful technologies and unreasonable applications of them. The intrinsic economic control has largely evaporated, partly because the tools use electronics and the components are produced in high volumes at low unit cost. Another reason is that the advertising and marketing sectors are highly sophisticated, already hold and exploit vast quantities of personal data, and are readily geared up to exploit yet more data.

Neither the oxymoronic notion of ‘business ethics’ nor the personal morality of executives in business and government act as any significant brake on the behaviours of corporations and governments, because they are very weak barriers, and they are readily rationalised away in the face of claims of enhanced efficiencies in, for example, marketing communications, fraud control, criminal justice and control over anti-social behaviour.

A further category of intrinsic control is ‘self-regulatory’ arrangements within relevant industry sectors. In 2010, for example, the Australian Mobile Telecommunications Association (AMTA) released industry guidelines to promote the privacy of people using LBS on mobile devices (AMTA, 2010). The guidelines were as follows:

1. Every LBS must be provided on an opt-in basis with a specific request from a user for the service

2. Every LBS must comply with all relevant privacy legislation

3. Every LBS must be designed to guard against consumers being located without their knowledge

4. Every LBS must allow consumers to maintain full control

5. Every LBS must enable customers to control who uses their location information and when that is appropriate, and be able to stop or suspend a service easily should they wish

The second point is a matter for parliaments, privacy oversight agencies and law enforcement agencies, and its inclusion in industry guidelines is for information only. The remainder, meanwhile, are at best ‘aspirational’, and at worst mere window-dressing. Codes of this nature are simply ignored by industry members. They are primarily a means to hold off the imposition of actual regulatory measures. Occasional short-term constraints may arise from flurries of media attention, but the ‘responsible’ organisations escape by suggesting that bad behaviour was limited to a few ‘cowboy’ organisations or was a one-time error that will not be repeated.

A case study of the industry self-regulation is provided by the Biometrics Code issued by the misleadingly named Australian industry-and-users association, the Biometrics ‘Institute’ (BI, 2004). During the period 2009–2012, the privacy advocacy organisation, the Australian Privacy Foundation (APF), submitted to the Privacy Commissioner on multiple occasions that the Code failed to meet the stipulated requirements and under the Commissioner's own Rules had to be de-registered. The Code never had more than five subscribers (out of a base of well over 100 members – which was itself only a sub-set of organisations active in the area), and had no signatories among the major biometrics vendors or users, because all five subscribers were small organisations or consultants. In addition, none of the subscribers appear to have ever provided a link to the Code on their websites or in their Privacy Policy Statements (APF, 2012).

The Commissioner finally ended the farce in April 2012, citing the “low numbers of subscribers”, but avoided its responsibilities by permitting the ‘Institute’ to “request” revocation, over two years after the APF had made the same request (OAIC, 2012). The case represents an object lesson in the vacuousness of self-regulation and the business friendliness of a captive privacy oversight agency.

If economics, morality and industry sector politics are inadequate, perhaps competition and organisational self-interest might work. On the other hand, repeated proposals that privacy is a strategic factor for corporations and government agencies have fallen on stony ground (Clarke, 19962006b).

The public can endeavour to exercise countervailing power against privacy-invasive practices. On the other hand, individuals acting alone are of little or no consequence to organisations that are intent on the application of location surveillance. Moreover, consumer organisations lack funding, professionalism and reach, and only occasionally attract sufficient media attention to force any meaningful responses from organisations deploying surveillance technologies.

Individuals may have direct surveillance countermeasures available to them, but relatively few people have the combination of motivation, technical competence and persistence to overcome lethargy and the natural human desire to believe that the institutions surrounding them are benign. In addition, some government agencies, corporations and (increasingly prevalent) public–private partnerships seek to deny anonymity, pseudonymity and multiple identities, and to impose so-called ‘real name’ policies, for example as a solution to the imagined epidemics of cyber-bullying, hate speech and child pornography. Individuals who use cryptography and other obfuscation techniques have to overcome the endeavours of business and government to stigmatise them as criminals with ‘something to hide’.

6.2. Legal controls

It is clear that natural or intrinsic controls have been utter failures in privacy matters generally, and will be in locational privacy matters as well. That leaves legal safeguards for personal freedoms as the sole protection. There are enormous differences among domestic laws relating to location surveillance. This section accordingly limits itself to generalities and examples.

Privacy laws are (with some qualifications, mainly in Europe) very weak instruments. Even where public servants and parliaments have an actual intention to protect privacy, rather than merely to overcome public concerns by passing placebo statutes, the draft Bills are countered by strong lobbying by government agencies and industry, to the extent that measures that were originally portrayed as being privacy-protective reach the statute books as authority for privacy breaches and surveillance (Clarke, 2000).

Privacy laws, once passed, are continually eroded by exceptions built into subsequent legislation, and by technological capabilities that were not contemplated when the laws were passed. In most countries, location privacy has yet to be specifically addressed in legislation. Even where it is encompassed by human rights and privacy laws, the coverage is generally imprecise and ambiguous. More direct and specific regulation may exist, however. In Australia, for example, the Telecommunications (Interception and Access) Act and the Surveillance Devices Act define and criminalise inappropriate interception and access, use, communication and publication of location information that is obtained from mobile device traffic (AG, 2005). On the other hand, when Google Inc. intercepted wi-fi signals and recorded the data that they contained, the Privacy Commissioner absolved the company (Riley, 2010), and the Australian Federal Police refused to prosecute despite the action – whether it was intentional, ‘inadvertent’ or merely plausibly deniable – being a clear breach of the criminal law (Moses, 2010Stilgherrian, 2012).

The European Union determined a decade ago that location data that is identifiable to individuals is to some extent at least subject to existing data protection laws (EU, 2002). However, the wording of that so-called ‘e-Privacy Directive’ countenances the collection of “location data which are more precise than is necessary for the transmission of communications”, without clear controls over the justification, proportionality and transparency of that collection (para. 35). In addition, the e-Privacy Directive only applies to telecommunications service-providers, not to other organisations that acquire location and tracking data. King and Jessen (2010) discuss various gaps in the protective regimes in Europe.

The EU's Advisory Body (essentially a Committee of European Data Protection Commissioners) has issued an Opinion that mobile location data is generally capable of being associated with a person, and hence is personal data, and hence is subject to the EU Directive of 1995 and national laws that implement that Directive (Art. 29, 2011). Consent is considered to be generally necessary, and that consent must be informed, and sufficiently granular (p. 13–8).

It is unclear, however, to what extent this Opinion has actually caused, and will in the future cause, organisations that collect, store, use and disclose location data to change their practices. This uncertainty exists in respect of national security, law enforcement and social control agencies, which have, or which can arrange, legal authority that overrides data protection laws. It also applies to non-government organisations of all kinds, which can take advantage of exceptions, exemptions, loopholes, non-obviousness, obfuscation, unenforceability within each particular jurisdiction, and extra-jurisdictionality, to operate in ways that are in apparent breach of the Opinion.

Legal authorities for privacy-invasions are in a great many cases vague rather than precise, and in many jurisdictions power in relation to specific decisions is delegated to a LEA (in such forms as self-written ‘warrants’), or even a social control agency (in the form of demand-powers), rather than requiring a decision by a judicial officer based on evidence provided by the applicant.

Citizens in many countries are subject to more or less legitimate surveillance of various degrees and orders of granularity, by their government, in the name of law enforcement and national security. However, many Parliaments have granted powers to national security agencies to use location technology to track citizens and to intercept telecommunications. Moreover, many Parliaments have failed the public by permitting a warrant to be signed by a Minister, or even a public servant, rather than a judicial officer (Jay, 1999). Worse still, it appears that these already gross breaches of the principle of a free society are in effect being extended to the authorisation of a private organisation to track mobiles of ordinary citizens because it may lead to better services planning, or more efficient advertising and marketing (Collier, 2011a).

Data protection legislation in all countries evidences massive weaknesses. There are manifold exemptions and exceptions, and there are intentional and accidental exclusions, for example through limitations in the definitions of ‘identified’ and ‘personal data’. Even the much vaunted European laws fail to cope with extraterritoriality and are largely ignored by US-based service-providers. They are also focused exclusively on data, leaving large gaps in safeguards for physical, communications and behavioural privacy.

Meanwhile, a vast amount of abuse of personal data is achieved through the freedom of corporations and government agencies to pretend that Terms imposed on consumers and citizens without the scope to reject them are somehow the subject of informed and freely given consent. For example, petrol stations, supermarkets and many government agencies pretend that walking past signs saying ‘area subject to CCTV’ represents consent to gather, transmit, record, store, use and disclose data. The same approach is being adopted in relation to highly sensitive location data, and much vaunted data protection laws are simply subverted by the mirage of consent.

At least notices such as ‘you are now being watched’ or ‘smile, you are being recorded’ inform customers that they are under observation. On the other hand, people are generally oblivious to the fact that their mobile subscriber identity is transmitted from their mobile phone and multilaterated to yield a reasonably precise location in a shopping mall (Collier, 2011a,b,c). Further, there is no meaningful sense in which they can be claimed to have consented to providing location data to a third party, in this case a location service-provider with whom they have never had contact. And the emergent combination of MDS with CCTV sources becomes a pervasive view of the person, an ‘über’ view, providing a set of über-analytics to – at this stage – shopping complex owners and their constituents.

What rights do employees have if such a system were instituted in an employment setting (Michael and Rose, 2007, p. 252–3)? Are workplace surveillance laws in place that would protect employees from constant monitoring (Stern, 2007)? A similar problem applies to people at airports, or on hospital, university, industrial or government campuses. No social contract has been entered into between the parties, rendering the subscriber powerless.

Since the collapse of the Technology Assessment movement, technological deployment proceeds unimpeded, and public risks are addressed only after they have emerged and the clamour of concern has risen to a crescendo. A reactive force is at play, rather than proactive measures being taken to ensure avoidance or mitigation of potential privacy breaches (Michael et al., 2011). In Australia, for example, safeguards for location surveillance exist at best incidentally, in provisions under separate legislative regimes and in separate jurisdictions, and at worst not at all. No overarching framework exists to provide consistency among the laws. This causes confusion and inevitably results in inadequate protections (ALRC, 2008).

6.3. Prospective legal controls

Various learned studies have been conducted, but gather dust. In Australia, the three major law reform commissions have all reported, and all have been ignored by the legislatures (NSWLRC, 2005ALRC, 2008VLRC, 2010).

One critical need is for the fundamental principle to be recovered, to the effect that the handling of personal data requires either consent or legal authority. Consent is meaningless as a control over unreasonable behaviour, however, unless it satisfies a number of key conditions. It must be informed, it must be freely given, and it must be sufficiently granular, not bundled (Clarke, 2002). In a great many of the circumstances in which organisations are claiming to have consent to gather, store, use and disclose location data, the consumer does not appreciate what the scope of handling is that the service-provider is authorising themselves to perform; the Terms are imposed by the service-provider and may even be varied or completely re-written without consultation, a period of notice or even any notice at all; and consent is bundled rather than the individual being able to construct a pattern of consents and denials that suit their personal needs. Discussions all too frequently focus on the specifically-US notion of ‘opt-out’ (or ‘presumed consent’), with consent debased to ‘opt-in’, and deprecated as inefficient and business-unfriendly.

Recently, some very weak proposals have been put forward, primarily in the USA. In 2011, for example, two US Senators proposed a Location Privacy Protection Bill (Cheng, 2011). An organisation that collected location data from mobile or wireless data devices would have to state explicitly in their privacy policies what was being collected, in plain English. This would represent only a partial implementation of the already very weak 2006 recommendation of the Internet Engineering Task Force for Geographic Location/Privacy (IETF GEOPRIV) working group, which decided that technical systems should include ‘Fair Information Practices’ (FIPs) to defend against harms associated with the use of location technologies (EPIC, 2006). FIPs, however, is itself only a highly cut-down version of effective privacy protections, and the Bill proposes only a small fraction of FIPs. It would be close to worthless to consumers, and close to legislative authorisation for highly privacy-invasive actions by organisations.

Two other US senators tabled a GPS Bill, nominally intended to “balance the needs of Americans' privacy protections with the legitimate needs of law enforcement, and maintains emergency exceptions” (Anderson, 2011). The scope is very narrow – next would have to come the Wi-Fi Act, the A-GPS Act, etc. That approach is obviously unviable in the longer term as new innovations emerge. Effective legislation must have appropriate generality rather than excessive technology-specificity, and should be based on semantics not syntax. Yet worse, these Bills would provide legal authorisation for grossly privacy-invasive location and tracking. IETF engineers, and now Congressmen, want to compromise human rights and increase the imbalance of power between business and consumers.

7. Conclusions

Mobile device location technologies and their applications are enabling surveillance, and producing an enormous leap in intrusions into data privacy and into privacy of the person, privacy of personal communications, and privacy of personal behaviour.

Existing privacy laws are entirely incapable of protecting consumers and citizens against the onslaught. Even where consent is claimed, it generally fails the tests of being informed, freely given and granular.

There is an urgent need for outcries from oversight agencies, and responses from legislatures. Individual countries can provide some degree of protection, but the extra-territorial nature of so much of the private sector, and the use of corporate havens, in particular the USA, mean that multilateral action is essential in order to overcome the excesses arising from the US laissez fairetraditions.

One approach to the problem would be location privacy protection legislation, although it would need to embody the complete suite of protections rather than the mere notification that the technology breaches privacy. An alternative approach is amendment of the current privacy legislation and other anti-terrorism legislation in order to create appropriate regulatory provisions, and close the gaps that LBS providers are exploiting (Koppel, 2010).

The chimeras of self-regulation, and the unenforceability of guidelines, are not safeguards. Sensitive data like location information must be subject to actual, enforced protections, with guidelines and codes no longer used as a substitute, but merely playing a supporting role. Unless substantial protections for personal location information are enacted and enforced, there will be an epidemic of unjustified, disproportionate and covert surveillance, conducted by government and business, and even by citizens (Gillespie, 2009Abbas et al., 2011).

Acknowledgements

A preliminary version of the analysis presented in this paper appeared in the November 2011 edition of Precedent, the journal of the Lawyers Alliance. The article has been significantly updated as a result of comments provided by the referees and editor.

References

R. Abbas, The social and behavioural implications of location-based services: an observational study of users, Journal of Location Based Services, 5 (December 2011), pp. 3-4

R. Abbas, K. Michael, M.G. Michael, A. Aloudat, Emerging forms of covert surveillance using GPS-enabled devices, Journal of Cases on Information Technology, 13 (2) (2011), pp. 19-33

AG, What the government is doing: Surveillance Device Act 2004, Australian Government (25 May 2005) at http://www.ag.gov.au/agd/www/nationalsecurity.nsf/AllDocs/9B1F97B59105AEE6CA25700C0014CAF5?OpenDocument

ALRC, For your information: Australian privacy law and practice, (ALRC report 108), Australian Government (2008), 2, p. 1409–10, http://www.alrc.gov.au.ezproxy.uow.edu.au/publications/report-108

AMTA, New mobile telecommunications industry guidelines and consumer tips set benchmark for location based services, Australian Mobile Telecommunications Association (2010) at http://www.amta.org.au/articles/New.mobile.telecommunications.industry.guidelines.and.consumer.tips.set.benchmark.for.Location.Based.Services

N. Anderson, Bipartisan bill would end government's warrantless GPS tracking, Ars Technica (June 2011) at http://arstechnica.com/tech-policy/news/2011/06/bipartisan-bill-would-end-governments-warrantless-gps-tracking.ars

APF Revocation of the biometrics industry code, Australian Privacy Foundation (March 2012) at http://www.privacy.org.au/Papers/OAIC-BiomCodeRevoc-120321.pdf

B. Arnold, Privacy guide, Caslon Analytics (May 2008), at http://www.caslon.com.au/privacyguide19.htm

Art. 29, Opinion 13/2011 on geolocation services on smart mobile devices, Article 29 Data Protection Working Party, 881/11/EN WP 185, at http://ec.europa.eu/justice/policies/privacy/docs/wpdocs/2011/wp185_en.pdf (16 May 2011)

BI Privacy code, Biometrics Institute, Sydney (April 2004) at http://web.archive.org/web/20050424120627/http://www.biometricsinstitute.org/displaycommon.cfm?an=1&subarticlenbr=8

A.J. Blumberg, P. EckersleyOn locational privacy, and how to avoid losing it forever, Electronic Frontier Foundation (August 2009), at https://www.eff.org/wp/locational-privacy

S. Bronitt, Regulating covert policing methods: from reactive to proactive models of admissibility, S. Bronitt, C. Harfield, K. Michael (Eds.), The social implications of covert policing (2010), pp. 9-14

J. Cheng, Franken's location-privacy bill would close mobile-tracking ‘loopholes’, Wired (17 June 2011), at http://www.wired.com/epicenter/2011/06/franken-location-loopholes/

K. Chetty, G.E. Smith, K. Woodbridge, Through-the-wall sensing of personnel using passive bistatic WiFi radar at standoff distances, IEEE Transactions on Geoscience and Remote Sensing, 50 (4) (April 2012), pp. 1218-1226

R. Clarke, Information technology and dataveillance, Communications of the ACM, 31 (5) (May 1988), pp. 498-512, at http://www.rogerclarke.com/DV/CACM88.html

R. Clarke, The digital persona and its application to data surveillance, The Information Society, 10 (2) (June 1994), pp. 77-92, at http://www.rogerclarke.com/DV/DigPersona.html

Clarke R. Privacy and dataveillance, and organisational strategy. In: Proc. I.S. Audit & Control Association (EDPAC'96), Perth, Western Australia; May 1996, at http://www.rogerclarke.com/DV/PStrat.html.

R. Clarke, Submission to the Commonwealth Attorney-General re: ‘a privacy scheme for the private sector: release of key provisions’ of 14 December 1999, Xamax Consultancy Pty Ltd (January 2000) at http://www.anu.edu.au/people/Roger.Clarke/DV/PAPSSub0001.html

R. Clarke, Person-location and person-tracking: technologies, risks and policy implications, Information Technology & People, 14 (2) (Summer 2001), pp. 206-231, at http://www.rogerclarke.com/DV/PLT.html

Clarke R. e-Consent: a critical element of trust in e-business. In: Proc. 15th Bled electronic commerce conference, Bled, Slovenia; June 2002, at http://www.rogerclarke.com/EC/eConsent.html.

R. Clarke, What's ‘privacy’? Xamax Consultancy Pty Ltd (2006), August 2006, at http://www.rogerclarke.com/DV/Privacy.html

R. Clarke, Make privacy a strategic factor – the why and the how, Cutter IT Journal, 19 (11) (2006), at http://www.rogerclarke.com/DV/APBD-0609.html

R. Clarke, Dissidentity: the political dimension of identity and privacy, Identity in the Information Society, 1 (1) (December 2008), pp. 221-228, at http://www.rogerclarke.com/DV/Dissidentity.html

Clarke R. The covert implementation of mass vehicle surveillance in Australia. In: Proc 4th workshop on the social implications of national security: covert policing, April 2009, ANU, Canberra; 2009a, at http://www.rogerclarke.com/DV/ANPR-Surv.html.

Clarke R. A sufficiently rich model of (id)entity, authentication and authorisation. In: Proc. IDIS 2009 – the 2nd multidisciplinary workshop on identity in the Information Society, LSE, 5 June 2009; 2009b, at http://www.rogerclarke.com/ID/IdModel-090605.html.

R. Clarke, A framework for surveillance analysis, Xamax Consultancy Pty Ltd (2009), August 2009, at http://www.rogerclarke.com/DV/FSA.html

R. Clarke, What is überveillance? (And what should be done about it?) IEEE Technology and Society, 29 (2) (Summer 2010), pp. 17-25, at http://www.rogerclarke.com/DV/RNSA07.html

Clarke R. The cloudy future of consumer computing. In: Proc. 24th Bled eConference; June 2011, at http://www.rogerclarke.com/EC/CCC.html.

R. Clarke, M. Wigan, You are where you've been: the privacy implications of location and tracking technologies, Journal of Location Based Services, 5 (3–4) (December 2011), pp. 138-155, http://www.rogerclarke.com/DV/YAWYB-CWP.html

E.B. Cleff, Implementing the legal criteria of meaningful consent in the concept of mobile advertising, Computer Law & Security Review, 23 (2) (2007), pp. 262-269

E.B. Cleff, Effective approaches to regulate mobile advertising: moving towards a coordinated legal, self-regulatory and technical response, Computer Law & Security Review, 26 (2) (2010), pp. 158-169

K. Collier, Stores spy on shoppers, Herald Sun (2011), 12 October 2011, at http://www.heraldsun.com.au/news/more-news/stores-spy-on-shoppers/story-fn7x8me2-1226164244739

K. Collier, Shopping centres' Big Brother plan to track customers, Herald Sun (2011), 14 October 2011, at http://www.heraldsun.com.au/news/more-news/shopping-centres-big-brother-plan-to-track-customers/story-fn7x8me2-1226166191503

K. Collier, ‘Creepy’ path intelligence retail technology tracks shoppers, news.com.au (2011), 14 October 2011, at http://www.news.com.au/money/creepy-retail-technology-tracks-shoppers/story-e6frfmci-1226166413071

F. Dahunsi, B. Dwolatzky, An empirical investigation of the accuracy of location-based services in South Africa, Journal of Location Based Services, 6 (1) (March 2012), pp. 22-34

J. Dobson, P. Fisher, Geoslavery, IEEE Technology and Society, 22 (2003), pp. 47-52, cited in Raper et al. (2007)

Economist, Vehicle data recorders – watching your driving, The Economist (23 June 2012), at http://www.economist.com/node/21557309

J. Edwards, Apple has quietly started tracking iphone users again, and it's tricky to opt out, Business Insider (11 October 2012) at http://www.businessinsider.com/ifa-apples-iphone-tracking-in-ios-6-2012-10

EPIC, Privacy and human rights report 2006, Electronic Privacy Information Center, WorldLII (2006) at http://www.worldlii.org.ezproxy.uow.edu.au/int/journals/EPICPrivHR/2006/PHR2006-Location.html

EPIC, Investigations of Google street view, Electronic Privacy Information Center (2012), at http://epic.org/privacy/streetview/

EU Directive 2002/58/EC of the European Parliament and of the Council of 12 July 2002 concerning the processing of personal data and the protection of privacy in the electronic communications sector (Directive on privacy and electronic communications)

Official Journal, L 201 (2002), 31/07/2002 P. 0037-0047, European Commission, at http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32002L0058:en:HTML

J. Figueiras, S. Frattasi, Mobile positioning and tracking: from conventional to cooperative techniques, Wiley (2010)

S.J. Fusco, R. Abbas, K. Michael, A. Aloudat, Location-based social networking and its impact on trust in relationships, IEEE Technology and Society Magazine, 31 (2) (Summer 2012), pp. 39-50, at http://works.bepress.com.ezproxy.uow.edu.au/cgi/viewcontent.cgi?article=1326&context=kmichael

Gallagher T et al. Trials of commercial Wi-Fi positioning systems for indoor and urban canyons. In: Proc. IGNSS symposium, Queensland; 1–3 December 2009, cited in Zandbergen (2012).

J.S. Ganz, It's already public: why federal officers should not need warrants to use GPS vehicle tracking devices, Journal of Criminal Law and Criminology, 95 (4) (Summer 2005), pp. 1325-1337

A.A. Gillespie, Covert surveillance, human rights and the law, Irish Criminal Law Journal, 19 (3) (August 2009), pp. 71-79

IBM, IBM smart surveillance system (previous PeopleVision project, IBM Research (30 October 2011), at http://www.research.ibm.com.ezproxy.uow.edu.au/peoplevision/

D.M. Jay, Use of covert surveillance obtained by search warrant, Australian Law Journal, 73 (1) (Jan 1999), pp. 34-36

N.J. King, P.W. Jessen, Profiling the mobile customer – privacy concerns when behavioural advertisers target mobile phones, Computer Law & Security Review, 26 (5) (2010), pp. 455-478, and 2010; 26(6): 595–612

A. Koppel, Warranting a warrant: fourth amendment concerns raised by law enforcement's warrantless use of GPS and cellular phone tracking, University of Miami Law Review, 64 (3) (April 2010), pp. 1061-1089

P. Lewis, Fears over privacy as police expand surveillance project, The Guardian (15 September 2008) at http://www.guardian.co.uk/uk/2008/sep/15/civilliberties.police

B. van Loenen, J. Zevenbergen, J. de JongBalancing location privacy with national security: a comparative analysis of three countries through the balancing framework of the European court of human rights, N.J. Patten, et al. (Eds.), National security: institutional approaches, Nova Science Publishers (2009), [chapter 2]

M. McGuire, K.N. Plataniotis, A.N. Venetsanopoulos, Data fusion of power and time measurements for mobile terminal location, IEEE Transaction on Mobile Computing, 4 (2005), pp. 142-153, cited in Raper et al. (2007)

S. Mann, J. Nolan, B. Wellman, Sousveillance: inventing and using wearable computing devices for data collection in surveillance environments, Surveillance & Society, 1 (3) (June 2003), pp. 331-355, at http://www.surveillance-and-society.org/articles1(3)/sousveillance.pdf

Mautz R. Overview of indoor positioning technologies. Keynote. In: Proc. IPIN'2011, Guimaraes; September 2011, at http://www.geometh.ethz.ch/people/.../IPIN_Keynote_Mautz_2011.pdf.

D. Mery, The mobile phone as self-inflicted surveillance – and if you don't have one, what have you got to hide? The Register (10 April 2009) at http://www.theregister.co.uk/2009/04/10/mobile_phone_tracking/

Michael and Michael, 2007, K. Michael, M.G. Michael, From dataveillance to überveillance and the Realpolitik of the Transparent Society, University of Wollongong (2007) at http://works.bepress.com.ezproxy.uow.edu.au/kmichael/51

K. Michael, M.G. Michael, Innovative automatic identification and location-based services: from bar codes to chip implants, IGI Global (2009)

M.G. Michael, K. Michael, Towards a state of uberveillance, IEEE Technology and Society Magazine, 29 (2) (Summer 2010), pp. 9-16, at, http://works.bepress.com.ezproxy.uow.edu.au/kmichael/187

Michael K, McNamee A, Michael MG, Tootell H., Location-based intelligence – modeling behavior in humans using GPS. In: Proc. int'l symposium on technology and society, New York, 8–11 June 2006; 2006a, at http://ro.uow.edu.au/cgi/viewcontent.cgi?article=1384&context=infopapers.

Michael K, McNamee A, Michael MG. The emerging ethics of humancentric GPS tracking and monitoring. In: Proc. int'l conf. on mobile business, Copenhagen, Denmark. IEEE Computer Society; 2006b, at http://ro.uow.edu.au/cgi/viewcontent.cgi?article=1384&context=infopapers.

M.G. Michael, S.J. Fusco, K. Michael, A research note on ethics in the emerging age of uberveillance, Computer Communications, 31 (6) (2008), pp. 1192-1199, at http://works.bepress.com.ezproxy.uow.edu.au/kmichael/32/

Michael and Masters, 2006, K. Michael, A. Masters, Realized applications of positioning technologies in defense intelligence, H. Hussein Abbass, D. Essam (Eds.), Applications of information systems to homeland security and defense, Idea Group Publishing (2006), at http://works.bepress.com.ezproxy.uow.edu.au/kmichael/2

K. Michael, G. Rose, Human tracking technology in mutual legal assistance and police inter-state cooperation in international crimes, K. Michael, M.G. Michael (Eds.), From dataveillance to überveillance and the realpolitik of the transparent society. 1st ed, University of Wollongong, Wollongong (2007), pp. 241-256.

K. Michael, G. Roussos, G.Q. Huang, R. Gadh, A. Chattopadhyay, S.Prabhu, et al.Planetary-scale RFID services in an age of uberveillance, Proceedings of the IEEE, 98 (9) (2010), pp. 1663-1671

K. Michael, M.G. Michael, R. Abbas, The importance of scenarios in the prediction of the social implications of emerging technologies and services, Journal of Cases on Information Technology (JCIT) 13.2 (2011), pp. i-vii

A. Moses, Google escapes criminal charges for Wi-Fi snooping, The Sydney Morning Herald (6 December 2010) at http://www.smh.com.au/technology/security/google-escapes-criminal-charges-for-wifi-snooping-20101206-18lot.html

NSWLRC Surveillance, Report 108, NSW Law Reform Commission (2005) at http://www.lawlink.nsw.gov.au/lawlink/lrc/ll_lrc.nsf/pages/LRC_r108toc

OAIC. Office of the Australian Information Commissioner; April 2012, at http://www.comlaw.gov.au/Details/F2012L00869/Explanatory%20Statement/Text.

A.A. Otterberg, Note: GPS tracking technology: the case for revisiting Knotts and shifting the Supreme Court's theory of the public space under the fourth amendment, Boston College Law Review, 46 (2005) (2005), pp. 661-704

C. Parenti, The soft cage: surveillance in America from slavery to the war on terror, Basic Books (2003)

PI, Our commitment to privacy, Path Intelligence (2010), heading changed in late 2012 to ‘privacy by design’, at http://www.pathintelligence.com/en/products/footpath/privacy

PI, FootPath technology, Path Intelligence (2010) at http://www.pathintelligence.com/en/products/footpath/footpath-technology

PI Retail, Path Intelligence (2012), at http://www.pathintelligence.com/en/industries/retail

J. Raper, G. Gartner, H. Karimi, C. Rizos, A critical evaluation of location based services and their potential, Journal of Location Based Services, 1 (1) (2007), pp. 5-45

J. Raper, G. Gartner, H. Karimi, C. Rizos, Applications of location-based services: a selected review, Journal of Location Based Services, 1 (2) (2007), pp. 89-111

RE IEEE 802.11 standards tutorial, Radio-Electronics.com (2010), apparently of 2010, at http://www.radio-electronics.com/info/wireless/wi-fi/ieee-802-11-standards-tutorial.php

RE WiMAX IEEE 802.16 technology tutorial, Radio-Electronics.com (2010), apparently of 2010, at http://www.radio-electronics.com/info/wireless/wimax/wimax.php

RE Assisted GPS, A-GPS, Radio-Electronics.com (2012) apparently of 2012, at http://www.radio-electronics.com/info/cellulartelecomms/location_services/assisted_gps.php

Renegar BD, Michael K, Michael MG. Privacy, value and control issues in four mobile business applications. In: Proc. 7th int'l conf. on mobile business; 2008. p. 30–40.

J. Riley, Gov't ‘travesty’ in Google privacy case, ITWire, 20 (Wednesday 3 November 2010), p. 44, at http://www.itwire.com/it-policy-news/regulation/42898-govt-travesty-in-google-privacy-case

I.J. Samuel, Warrantless location tracking, New York University Law Review, 83 (2008), pp. 1324-1352

SHW Skyhook location performance at http://www.skyhookwireless.com/location-technology/performance.php (2012)

Skyhook. (2012). Website entries, including ‘frequently asked questions’ at http://www.skyhookwireless.com/whoweare/faq.php, ‘privacy policy’ at http://www.skyhookwireless.com/whoweare/privacypolicy.php and ‘location privacy’ at http://www.skyhookwireless.com/whoweare/privacy.php.

C. Song, Z. Qu, N. Blumm, A.-L. Barabási, Limits of predictability in human mobility, Science, 327 (5968) (2010), pp. 1018-1021.

A. Stern, Man fired thanks to GPS tracking, Center Networks (31 August 2007), at http://www.centernetworks.com/man-fired-thanks-to-gps-tracking

Stilgherrian, Forget government data retention, Google has you wired, Crikey (2 October 2012), at http://www.crikey.com.au/2012/10/02/forget-government-data-retention-google-has-you-wired/

USGovGPS accuracy, National Coordination Office for Space-Based Positioning, Navigation, and Timing(February 2012), at http://www.gps.gov/systems/gps/performance/accuracy/

VLRC, Surveillance in public spaces, Victorian Law Reform Commission (March 2010), Final report 18, at http://www.lawreform.vic.gov.au/wps/wcm/connect/justlib/Law+Reform/resources/3/6/36418680438a4b4eacc0fd34222e6833/Surveillance_final_report.pdf

D. Wright, M. Friedewald, S. Gutwirth, M. Langheinrich, E. Mordini, R.Bellanova, et al.Sorting out smart surveillance, Computer Law & Security Review, 26 (4) (2010), pp. 343-354

P.A. Zandbergen, Comparison of WiFi positioning on two mobile devices, Journal of Location Based Services, 6 (1) (March 2012), pp. 35-50

Keywords: Location-based systems (LBS), Cellular mobile, Wireless LAN, GPS, Mobile device signatures (MDS), Privacy, Surveillance, Überveillance

Citation: Katina Michael and Roger Clarke, "Location and tracking of mobile devices: Überveillance stalks the streets", Computer Law & Security Review, Vol. 29, No. 3, June 2013, pp. 216-228, DOI: https://doi.org/10.1016/j.clsr.2013.03.004

RFID-Enabled Inventory Control Optimization

Abstract

This study examines the impact of radio-frequency identification (RFID) technology on the inventory control practices of a small-to-medium retailer using a proof of concept (PoC) approach. The exploratory study was conducted using a single case study of a hardware retailer stocking 5000 product lines provided by 110 active suppliers. To analyze the present mode of operation, procedural documents, semi-structured interviews and a participant observation was conducted. The basis for the proof of concept was a future mode of operation using a quasi-experimental design. Results indicate that in a small-to-medium retail environment, RFID technology could act as a loss prevention mechanism, an enabler for locating misplaced stock, and make a significant contribution to the overall improvement of the delivery process.

Section I

Introduction

Radio-frequency identification (RFID), which is defined as a wireless automatic identification and data capture (AIDC) technology [1], is increasingly considered by many scholars as the “missing link” in the supply chain management [2], [3]. For example, the technology could allow the identification of any tagged item in real-time in a given supply chain with minimum human intervention [4] [5] [6] [7]. When integrating into a firm's business processes [5], the RFID technology allows “any tagged entity to become a mobile, intelligent, communicating component of the organization's overall information infrastructure” (p. 88), thus improving supply chain information flow [8], [9] and supply chain efficiency [3]. A basic RFID system is composed of a tag containing a microprocessor, a reader and its antennas, and a computer equipped with a middleware program, in which business rules are configured to automate some decisions [10]. Despite the high potential of the technology as an enabler of the supply chain transformation, the current adoption rate is still fairly low mainly because many technological and business questions are still to be answered. In order to reduce this knowledge gap, this study draws on the current RFID agenda [5] to answer the following questions: What is the impact of RFID on loss prevention? What is the impact of RFID technology on the delivery process in a small-to-medium retailer store? How can RFID help to locate misplaced stock? How may the RFID reading rate be influenced by the physical characteristics of items? More precisely, the objective of this paper is to document the results of a proof of concept (PoC) that examines the impact of RFID on inventory control. The PoC consists of RFID simulations and re-engineered business processes that demonstrate whether the RFID technology can operate within the small-to-medium retail industry and illustrates the anticipated impact of RFID on business operations.

Section 2 presents related works. In section 3, the methodology used in this study, including all simulation of RFID enabled scenarios are presented. Finally, section 4 presents the discussion and conclusion.

Section II

Background and Context of the Study

The current study uses the proof of concept approach to assess the feasibility of RFID technology in a small-to-medium retail store. Most early studies on the feasibility of RFID technology have mostly been conducted using this approach or pilots projects (e.g. [11] [12] [13]). A proof of concept is used to illustrate whether a proposed system or application is likely to operate or function as expected [14]. Using, data from “Wal-Mart RFID-enabled stores” over a period of 29 weeks, the conclusion was reached that RFID-enabled stores were 63% more efficient in replenishing out-of-stocks than stores without RFID, thus leading to a reduction of out-of-stocks by 16% over that 29 week period [12]. In a more recent study, [15] examined data collected over a period of 23 weeks from eight test stores equipped with a new “RFID-based perpetual inventory adjustment tool” and a corresponding set of eight control stores (without RFID), and found that “RFID is making a difference. Understated perpetual inventory inaccuracy declined by about 13% in the test stores, relative to control stores, with no additional labour. Furthermore, manual adjustments declined in the test stores” (p. 55). Finally, the outcome of a study by [11], who used a PoC in a laboratory setting, was that process optimization can be achieved when the RFID technology is integrated in intra- and inter-organizational information systems applications. All these studies have been largely conducted in large firms, but very few of them are concerned with RFID adoption within small-to-medium firms.

Section III

Methodology

The research study documented in this paper involves a case examining a single small-to-medium retailer. A case study method has been employed as it is ideal for investigating contemporary events and is able to take into account a wide variety of evidence [16]. For this study data have been gathered through the collection of procedural documents, semistructured interviews and a participant observation. This paper presents the data collected from the semistructured interviews conducted with employees of the organization, as well as revealing the business process flows (through flowcharts) of the organization in order to determine whether RFID is a feasible automated data capture technology for small-to-medium retailers. An observational study was also conducted over a period of two weeks in 2007. A daily diary was kept by the participant and this data was analyzed together with full-length transcripts. A single small-to-medium hardware retailer is focused on in this paper in order to analyze and present inventory control practices.

3.1. Research design

As the main objective of the overall study is to improve our understanding of RFID impacts in the context of a small-to-medium retailer, the research design is clearly an exploratory research initiative. A case study method has been conducted as it is ideal for investigating contemporary events and is able to take into account a wide variety of evidence [16].

3.2. Research sites

The organization examined in this study is located on the south coast of New South Wales, approximately 128 kilometers from the centre of Sydney. The company employs ten staff including casuals and is classified as a small-to-medium hardware retailer. The current proprietors have operated the business since 2003. The premises of the retailer measures approximately 2000 square meters, with about 550 square meters of this area making up the internal shop floor. The shop floor is composed of four sheds, each with independent access. There are two small internal offices, one designed to deal with customer purchasing and point-of-sale (POS) transactions while the other is used by managers and bookkeepers for ordering, accounting and other administrative practices. The external perimeter of the organization is surrounded by an eight foot high barbed wire fence.

The retailer currently possesses between 400,000 worth of inventory which is kept on the premises. The inventory held by the organization is estimated to consist of 5000 product lines, which are provided by 110 active suppliers. Products and other inventory are stored or displayed before purchase inside the store or outside within the confines of the premises. Items and stock within the store are positioned based on the type of product as well as the supplier. Most items kept inside the store are also shelved on racks that measure 2.1m in height. The shop floor is divided into five separate areas that include general hardware, timber, gardening, cement and building supplies. Products stored outside are generally unaffected by environmental and weather conditions such as landscaping supplies, cement blocks, treated pine sleepers and sheets of steel reinforcing. Stock is usually delivered to the store packaged at pallet, crate, carton or item level.

The retailer provides many services to its customers primarily through the selling of hardware and other building related supplies. The organization provides a delivery service to its customers if they purchase products that are too large to be transported or products that they wish to be delivered on a certain day. Products are delivered to customers in one of the three vehicles the organization owns. A flat top truck is used for steel deliveries, a tip truck is used for landscaping supplies and a utility vehicle is used for general deliveries. The organization also has a front-end loader that it uses to load landscaping supplies on vehicles. The organization offers accounts for customers that purchase products frequently.

The retailer currently has limited Information Technology (IT) infrastructure and does not utilize a server, as the current operations of the business do not require a large volume storage device. The organization utilizes two desktop computers in their administration office that are primarily used to manage customer accounts through the software package MYOB Premier Version 10. At the end of each month, the organization uses the MYOB software to generate invoices which are sent out to account holding customers, requesting that they pay for the items they have purchased. The organization has another desktop computer which is used by employees to search a program that acts as an index of paint colors provided by different paint suppliers. All computers within the organization are able to access the Internet.

3.3. Data collection

For this study data have been gathered through the collection of procedural documents, semi-structured interviews and a participant observation. This paper presents the data collected from the semi-structured interviews conducted with employees of the organization, as well as revealing the business process flows (through flowcharts) of the organization in order to determine whether RFID is a feasible automated data capture technology for small-to-medium retailers. An observational study was also conducted over a period of two weeks in 2007. A daily diary was kept by the participant and this data was analyzed together with full-length transcripts. A single small-to-medium hardware retailer is focused on in this paper in order to analyze and present inventory control practices.

3.3.1. Interviews-interviewees

Insights into the current inventory control practices at the small-to-medium retailer are based on semi-structured interviews carried out on four employees of the organization. The roles and duties of these employees are documented in Table 1.

Table 1. EMPLOYEE ROLES AND DUTIES

Table 1. EMPLOYEE ROLES AND DUTIES

As can be seen from Table 1, employees of the organization have minimal job specialization, which reinforces [17] observations of small businesses. The proprietor/manager and proprietor/part-time manager are responsible for the overall running of the business whereas the store manager is specifically responsible for shop maintenance and management. The delivery truck driver is primarily responsible for making outbound deliveries. The store manager and delivery truck driver are answerable to both of the proprietor/managers.

3.3.2. Interview questions and the inventory cycle

Inventory control as defined by [18] involves “coordinating the purchasing, manufacturing and distribution functions to meet marketing needs”. Coordinating these functions requires many discrete activities including ordering stock or materials and shelving or putting it in the correct position so that customers have access to it. In this section, the inventory control process has been broken down so that the inventory practices of the small-to-medium retailer can be explored in greater detail. Figure 1 illustrates the inventory cycle. It should be noted that the inventory flow cycle is focused on the flow of raw materials to their finished state, while this inventory control cycle has been developed based on a retailer that sells finished goods (p. 21) [19].

Figure 1. The Inventory Cycle

Figure 1. The Inventory Cycle

 

As can be seen in Figure 1, customer demand triggers the ordering or re-ordering of stock. Stock then arrives at the retailer, where it is checked and sorted before being shelved in the correct position. Stock is then purchased by a customer and delivered by the retailer if necessary.

The inventory cycle demonstrated in Figure 1 was considered when developing questions for the semistructured interviews. The majority of the questions asked related to the six different processes that were identified in the inventory control cycle. There were a total of twenty-eight questions included in the original semi-structured interview protocol but additional probing sub-questions were asked where the respondent was able to expand their response due to their knowledge of operations. The questions covered the background of the company case, the role of the employee in the organization, questions related to the current mode of operation to gauge the current inventory control practices and set-up, and more speculative questions regarding the transition of the organization from a manual-based system to barcode and/or RFID. For instance the proprietor was asked:

Can you describe the process that you use to check that orders have been delivered with the correct contents?

  1. Do you keep any sort of record of how much stock you carry, either in physical or electronic form?
  2. How would you describe the theft prevention measures in your workplace?
  3. What triggers your organization to reorder or order stock?
  4. Are there any issues affecting your adoption of automated data capture technology?
  5. Do you think that RFID could be used within your business to improve inventory control?

The interview transcripts were analyzed using a qualitative approach and the findings were presented using a modular narrative style based on the steps in the inventory control cycle. The following sections summarize the findings of the semi-structured interviews.

3.3.3. Participant observation

A participant observation requires the researcher to become a direct participant in the social process being studied by becoming a member of an organization. The participant observation was carried out over a two week period with the intention of recording observations relating to the inventory control practices used within the small-to-medium retailer. This study utilizes an overt participant observation as members of the organization were already aware of the researcher's presence due to interviews being carried out at an earlier date. The overt approach was perceived to have had minimal influence on the behavior of the organization's members as they were informed that the purpose of the study was to examine inventory control practices of the retailer, not their personal behaviors. During the participant observation annotations and issues were documented through the use of a diary. Field notes were recorded during each day, and were formalized at the end of the day.

3.3.4. Procedural documentation

The small-to- medium retailer's procedural documents were used to complement the semi-structured interviews and participant observation. Documentary secondary data, such as an organization's communications, notes, and other policy and procedural documents have been examined.

Table 2. THE FOUR RFID-ENABLED SCENARIOS

Table 2. THE FOUR RFID-ENABLED SCENARIOS

Official documents, like procedural documents can be treated as unproblematic statements of how things are or were (p. 104) [20]. The procedural documents have been used as evidence to support the determination of the inventory control practices of the small-to-medium retailer. The interviews conducted, participant observation and the collection of procedural documents were combined to develop the business process flows of the organization. A narrative presentation is used to bring together participant observational data and interviewee responses.

3.4. Simulation of RFID-enabled scenarios

Eight simulations have been developed which are aimed at examining different aspects of inventory control and known RFID issues that have been documented in the literature. However, within the scope of this paper, we'll only present and discuss four RFID-enabled scenarios (Table 2): (i) RFID-enabled loss prevention, (ii) RFID-enabled delivery portal, (iii) RFID tag environment simulation and (iv) RFID-enabled locating misplaced stock.

The results of the simulations are documented qualitatively, discussing read rates as well as any other technical issues experienced in the following section.

3.4.1. RFID enabled-loss prevention simulation-method

Exhibit 1. An RFID armed entry/exit

Exhibit 1. An RFID armed entry/exit

A fixed RFID reader with one and then two antennas will be placed above or around the entry/exit of the store with the aim of identifying any tagged item or product that passes through the entry/exit. Items that have been tagged with RFID labels will be moved past the reader in order to determine if the tag is interrogated and identified successfully. The tagged product will be concealed by the participant carrying it so the effect of this can be gauged. Multiple items will also be carried out by the participant to test if the reader identifies multiple tagged items.

In the initial part of this simulation a fixed reader was set up with one antenna which was positioned above the entry/exit, 2.1 metres off the ground.

The antenna was orientated at a 45 degree angle, sloping inwards towards the interior of the store. The participant walked towards the entry/exit with an RFID tagged item held 1.5 meters off the ground. Five different items of stock were used in this simulation, each being RFID tagged in a different configuration. Two of the items had tags wrapped around them so the tag was overlapping itself, one item had its tag wrapped around it but was not overlapping, another item was labelled with a tag that was folded in half and the final item had a tag applied to it in a general flat configuration. The tagged items were passed through the RFID monitored entry/exit individually in plain view of the reader, then concealed under the jumper of the participant and finally all items were passed through the entry/exit simultaneously in a plastic basket.

The results revealed that items which had RFID tags wrapped around them and were overlapping could not be detected by the reader when passed through the entry/exit. It was also found that concealing items had an effect on whether they would read or not with a single concealed product being identified compared to the three tagged items which were identified when they were passed through the entry/exit in plain sight. Table 3 summarises the results of the simulation = read successfully, = not able to be read).

Figure 2. Configuration of the loss prevention portal

Figure 2. Configuration of the loss prevention portal

Once this simulation was carried out another antenna was attached to the fixed reader and a small portal was created to see whether it was more accurate to identify tagged products from side-on than from above. Figure 2 illustrates the configuration of the portal.

The participant once again walked through the doorway with items held 1.5 metres from the ground. The items that had RFID tags wrapped around so they overlapped were still not able to be read in this variation of the simulation, but three tagged items that had been concealed were identified compared to the one item identified in the previous variation. The range of the antennas was also tested with items being passed through the portal held above (1.8 metres from the ground) and below them (30 centimetres from the ground). The three tagged items that were identified initially were also read when they were passed above and below the antennas at the entry/exit to the store.

Table 3. LOSS PREVENTION SIMULATION RESULTS

Table 3. LOSS PREVENTION SIMULATION RESULTS

The results of this simulation revealed that RFID experienced poor to average read rates when implemented for loss prevention. It is perceived that if RFID was applied in the small-to-medium retailer for loss prevention purposes, theft may be reduced but the reliability of the technology could not be guaranteed; unless orientation issues are resolved and read rates are improved.

 

3.4.2. RFID-enabled delivery portal simulation-method

This simulation involves RFID tagged items being placed on a pallet then onto a delivery vehicle at the loading dock of the hardware store. A portal will be created at the loading dock, equipped with two antennas originating from an RFID reader which will be used to identify the products and stock that are moving in and out of the premises.

Exhibit 2. Tagged RFID products on pallet (top); the flat top truck being reversed into the loading dock (middle); the utility vehicle in the RFID portal (bottom)

Exhibit 2. Tagged RFID products on pallet (top); the flat top truck being reversed into the loading dock (middle); the utility vehicle in the RFID portal (bottom)

To test the RFID delivery portal, a flat top truck is reversed into the loading bay of the organisation. Seven products that are commonly delivered to or by the organisation are RFID tagged, including a wooden pallet which the items are placed on. The truck is reversed in and out of the loading bay on five occasions and the read rates are recorded each time.

Three of the tagged items including a piece of treated pine, a roll of foam joint and the pallet are successfully interrogated on each of the five times the truck is reversed.

A tagged piece of treated pine is also identified on the first and the last time the vehicle is backed into the loading dock.

The other three items on the truck are unable to be identified at all, most likely due to the back tray of the truck, sitting higher than the antenna (all the RFID tagged products on the truck were situated above the antenna).

Another vehicle, a utility that is used by the organisation to deliver products is then employed in the simulation with the same products and pallet being placed in the vehicle's tray. The tray of this vehicle is at a more suitable height for the RFID antennas, as it sits 80 centimetres off the ground. Exhibit 2 demonstrates the RFID portal with the utility vehicle reversed into the loading dock.

The read rates experienced when products were placed on the utility were far superior to those experienced when the flat top truck was employed, with read rates ranging from 71% to 100% of all items and products tagged. Table 4 reveals the read rates of the tagged items and products on the utility vehicle (= read successfully, = not able to be read). This simulation illustrated that if an RFID portal was constructed appropriately by considering the conditions and vehicle used by the small-to-medium retailer it could effectively monitor stock being delivered to the business and stock being delivered to customers of the business.

Table 4. READ RATES OF RFID TAGGED ITEMS ON THE UTILITY VEHICLE

Table 4. READ RATES OF RFID TAGGED ITEMS ON THE UTILITY VEHICLE

3.4.3. RFID tag environment simulation-method

This simulation involves trying to identify RFID tagged products of various compositions using the mobile RFID reader. Items composed of wood, metal, plastic, stone and those containing liquids were tagged and attempted to be read. Items left outside and exposed to the elements were also tagged and attempted to be read, along with other items that are stored in dirty manufacturing type environments.

Ten products composed of varying materials were RFID tagged and attempted to be read by the mobile RFID reader. The compositions of the ten items tagged varied greatly with some of them being made or packaged from metal, plastic, cardboard, paper, wood and stone. Some of the items such as the container of nails and the bag of cement were also dusty and dirty. The mobile RFID reader was used to make six attempts to read data from all of the tagged products individually. Table 5 reveals the results of the six attempts for each product (= read successfully, = not able to be read).

Exhibit 3. An RFID tagged treated pine sleeper (top); An RFID tagged pipe (bottom)

Exhibit 3. An RFID tagged treated pine sleeper (top); An RFID tagged pipe (bottom)

As can be seen in Table 5 all items could be read by the mobile reader, but objects made of metal took around 5 or 6 attempts to be read successfully. It should also be noted that dirty and dusty products were interrogated successfully by the reader on every attempt.

In order to further test the effect the environment had on the readability of tags, four items that were regularly kept outside were RFID tagged. These items included a treated pine sleeper, a stone paver, a bale of sugar cane mulch wrapped in plastic and a 6 metre length galvanised pipe (Exhibit 3).

Table 5. READ RATES OF THE ENVIRONMENT SIMULATION

Table 5. READ RATES OF THE ENVIRONMENT SIMULATION

After being tagged with RFID labels these items were left outside for five nights. It rained quite heavily over the time the items were left outside and upon examining the products and RFID tags after the fifth night had elapsed, they were saturated.

This however did not have any effect on the readability of tags, with all items being successfully identified in all six of the scans except for the metal item (the 6 metre length of galvanised pipe) which was only interrogated successfully on the sixth attempt.

To compare the robustness of RFID tags and barcodes, a cardboard box with a barcode imprinted on it in ink was also left outside over the same period as the RFID tagged items. Like the RFID tags and products the cardboard box was saturated after the fifth night outside. The barcode on the box was able to be scanned successfully, but when the researcher applied some friction to the barcode it was damaged. Once the barcode was damaged it could not be identified by the barcode reader. Unlike the barcode the RFID tags were not affected or damaged by friction in this simulation.

This simulation revealed that the readability of RFID tags was not affected when applied to products of varying compositions, except for products composed of metal which resulted in these products only being identified in about one out of six attempts. It was also revealed that RFID tags were able to function after being stored outdoors and exposed to the elements over five nights. To further test the robustness of RFID tags it is recommended that they are exposed to the same environmental conditions for longer periods of time in a future study.

3.4.4. RFID-enabled- locating misplaced stock simulation-method

An RFID tagged product will be positioned so that it can be read by an antenna attached to a fixed RFID reader. Once data have been read from the tagged item it will then be moved around the shop to another location so it is within range of another antenna. The results of this simulation will focus on the ‘tag reads’ at each of the antennas. After one tagged item has been tested the read rates of multiple items will be observed.

A fixed RFID reader was set up with two antennas situated 10 metres apart. RFID tagged items were initially positioned in front of an antenna then put on a trolley and moved outside the range of the antenna and into the range of a second antenna to simulate stock being misplaced within the retailer. Exhibit 4 shows RFID tagged products that have been moved past an antenna on a trolley.

Exhibit 4. RFID tagged cartons of nails within the read range of an antenna

Exhibit 4. RFID tagged cartons of nails within the read range of an antenna

A plastic 5 kilogram carton of galvanised bullet head nails was RFID tagged and moved from the read range of the first antenna to within the read range of the second antenna which resulted in it being detected by both antennas. Once a single RFID tagged carton was tested more were introduced to further examine the accuracy of the antennas. Table 6 illustrates the results of this simulation. It should be noted that in the table, tags which were identified by both antennas (at the first antenna prior to being ‘misplaced’ and or the second antenna after being ‘misplaced’) were recorded as being read successfully (✓=read successfully,= not able to be read).

Table 6. PRODUCTS IDENTIFIED IN THE LOCATING MISPLACED STOCK SIMULATION

Table 6. PRODUCTS IDENTIFIED IN THE LOCATING MISPLACED STOCK SIMULATION

The results revealed read rates ranging from 67% to 100% for the five tests conducted in this simulation. When products were placed on the trolley and transported between antennas they were placed in a random configuration which meant that the RFID tags applied to them were not presented to the reader in the same arrangement for each of the tagged cartons of nails.

It was noted that tagged cartons that were not detected when moved between antennas, had tags orientated perpendicular to them or had tags that were applied to the opposite side of products. Figure 3 illustrates where RFID tags were applied on products that were not identified by the antennas.

Figure 3. The position of RFID tags that were not identified

Figure 3. The position of RFID tags that were not identified

Apart from the orientation issues that were encountered, this simulation illustrated that RFID could be used within the small-to-medium retailer to monitor the positioning of products within the store. If RFID was employed in the store and appropriate backend software was developed it is highly likely that misplaced items that had been tagged within the store could be registered on the system, and found thereafter.

Section IV

Discussion and Conclusion

The simulations revealed that items with overlapping RFID tags wrapped around them could not be detected by the reader when they passed through the entry/exit. It was also found that concealing items had an effect on whether they would read or not with a single concealed product being identified, as compared to the three tagged items which were identified when passing through the entry/exit in plain sight. Moreover, the results showed that RFID experienced poor to average read rates when implemented for loss prevention. It is perceived that if RFID was applied in the small-to-medium retailer for loss prevention purposes, theft may be reduced but the reliability of the technology could not be guaranteed, unless orientation issues were resolved and the read rates improved. Also, if an RFID portal were constructed appropriately, taking into account the conditions and the vehicle used by the small-to-medium retailer, it could effectively monitor the stock being delivered to the business and the one delivered to the customers of the business. In addition, the study revealed that the readability of RFID tags was not affected when applied to products of varying compositions, except for metal products - which were identified only once on six attempts. Moreover, the RFID tags were able to function after being stored outdoors and exposed to the elements over five nights. These results provide strong support to previous studies on RFID technology [11], [12] and highlight the fact that RFID technology is mostly product driven, and therefore, the best performance of the system heavily depends on the type of product, the context of implementation, the level of tagging, etc.

Consequently, a scenario building, validation and demonstration of RFID-enabled process optimization is highly recommended prior to any large RFID technology deployment [13]. To our knowledge, this study is among the first studies to illustrate that RFID technology could be used within a small-to-medium retailer in real-life settings to monitor the positioning of products within the store, to help small-to-medium retailer prevent in-store stock losses, enhance delivery process and improve the process of locating misplaced stock within the store. Nevertheless, these findings are consistent with results of prior research by [15] at Wal-Mart stores, which are mainly large stores. Despite these encouraging results, further tests on the robustness of RFID tags should be conducted when they are exposed to the same environmental conditions for longer periods of time. Moreover, given that the more recent RFID tags have a tag reading accuracy of almost 100%, their use is highly recommended [21]. The study was conducted in a single store of a small-to-medium retailer situated almost at the last node of the retail supply chain, and therefore was not able to capture the network effects of RFID technology.

Therefore, further works need to be done to assess the impact of RFID technology at the supply chain level in a real-life setting and to develop different models of cost sharing between stakeholders involved in RFID-enabled projects.

References

1. S. Fosso Wamba, L. A. Lefebvre, Y. Bendavid, and É. Lefebvre, "Exploring the impact of RFID technology and the EPC network on mobile B2B eCommerce: a case study in the retail industry," International Journal of Production Economics (112:2), 2008, 614-629.

2. R. Roman and J. Donald, "Impact of RFID technology on supply chain management systems," in 19th Annual Conference of the National Advisory Committee on Computing Qualifications (NACCQ 2006) Wellington, New Zealand, 2006.

3. C. Loebbecke, J. Palmer, and C. Huyskens, "RFID's potential in the fashion industry: a case analysis," in 19th Bled eConference, eValues Bled, Slovenia, 2006.

4. C. Poirier and D. McCollum, RFID Strategic Implementation and ROI: a Practical Roadmap to Success. Florida: J. ROSS Publishing, 2006.

5. J. Curtin, R. J. Kauffman, and F. J. Riggins, "Making the most out of RFID technology: a research agenda for the study of the adoption, usage and impact of RFID," Information Technology and Management (8:2), 2007, 87-110.

6. N. Huber and K. Michael, "Minimizing product shrinkage across the supply chain using radio frequency identification: A case study on a major Australian retailer," in IEEE Computer Society of the Seventh International Conference on Mobile Business Toronto, Canada, 2007.

7. B. D. Renegar and K. Michael, "The RFID value proposition," in CollECTeR Iberoamérica Madrid, Spain, 2008.

8. F. J. Riggins and K. T. Slaughter, "The role of collective mental models in IOS adoption: opening the black box of rationality in RFID deployment," in Proceedings of the 39th Hawaii International Conference on System Sciences Hawaii, 2006.

9. S. Fosso Wamba and H. Boeck, "Enhancing information flow in a retail supply chain using RFID and the EPC network: a proof-of-concept approach," Journal of Theoretical and Applied Electronic Commerce Research (3:1), 2008, 92-105.

10. Z. Asif and M. Mandviwalla, "Integrating the supply chain with RFID: a technical and business analysis," Communications of the Association for Information Systems (15), 2005, 393-427.

11. Y. Bendavid, S. Fosso Wamba, and L. A. Lefebvre, "Proof of concept of an RFID-enabled supply chain in a B2B e-commerce environment," in The Eighth International Conference on Electronic Commerce (ICEC) Fredericton, New Brunswick, Canada, 2006, 564-568.

12. B. C. Hardgrave, M. Waller, and R. Miller, " Does RFID reduce out of stocks? a preliminary analysis," 2005.

13. S. Fosso Wamba, E. Lefebvre, Y. Bendavid, and L. A. Lefebvre, From automatic identification and data capture (AIDC) to "smart business process": a proof of concept integrating RFID: CRC Press, Taylor & Francis Group, 2008.

14. W. E. Solutions, "Appendix A: Glossary," 1996.

15. B. C. Hardgrave, J. Aloysius, and S. Goyal, "Does RFID improve inventory accuracy? a preliminary analysis," International Journal of RF Technologies: Research and Applications (11:1), 2009, 44-56.

16. R. K. Yin, Case Study Research: Design and Methods. Newbury Park, CA: Sage, 1994.

17. J. Diamond and G. Pintel, Retailing. Upper Saddle River: Prentice Hall, 1996.

18. T. Wild, Best Practice in Inventory Management. New York: John Wiley & Sons, 1997.

19. R. Tersine, Principles of Inventory and Material Management. Upper Saddle River: Prentice Hall, 1998.

20. P. Knight, Small-Scale Research. London: Sage, 2002.

21. M. H. M. News, "UHF Gen 2 RFID delivers 100% read accuracy for item tagging," 2009.

IEEE Keywords: Australia, Business process re-engineering, Hardware, Humans, Inventory control, Radio frequency, Radiofrequency identification, Supply chain management, Supply chains, Testing

INSPEC: optimisation, radiofrequency identification, retail data processing, small-to-medium enterprises, stock control, RFID-enabled inventory control optimization, delivery process, hardware retailer, participant observation, procedural documents, proof of concept approach, quasi experimental design, radio-frequency identification technology, semi structured interviews, small-to-medium retailer

Citation: Dane Hamilton, Katina Michael, Samuel Wamba, 2010, "RFID-Enabled Inventory Control Optimization: A Proof of Concept in a Small-to-Medium Retailer", 2010 43rd Hawaii International Conference on System Sciences (HICSS), Date of Conference: 5-8 Jan. 2010, DOI: 10.1109/HICSS.2010.473

Privacy-value-control harmonization for RFID adoption in retail

Abstract

Privacy concerns have, at least in part, impeded the adoption of radio frequency identification (RFID) in retail. The adoption of other automatic identification (auto-ID) applications shows that consumers often are willing to trade their privacy or their control of personal information against some value afforded by the application. In this paper, the interplay between privacy, value, and control is examined through a literature survey of four auto-ID applications: mobile phone, electronic toll collection, e-passports, and loyalty programs. The consumer value proposition for the use of RFID in retail is investigated through an online survey exploring end-user perceptions. The results of the survey are: 1) the customer value proposition has not been communicated well to customers; 2) privacy concerns are higher than other previously adopted applications despite similar privacy issues; and 3) harmonization of privacy, value, and control is likely to be achieved only after adoption, when customers will be educated through experience with the application.

Introduction

Over the past decade, organizations have aggressively pursued the use of radio frequency identifi- cation (RFID) as a means to better identify, control, and track stock throughout the supply chain. The linking of RFID, an automatic identification (autoID) and data collection technology, to consumer goods has resulted in widespread concern surrounding privacy issues. The mainstream media have been quick to expose these privacy concerns, with most articles focusing purely on the potential of the technology to track consumers without their knowledge or consent. Prior to 2004, this resulted in many major retail organizations around the world temporarily halting their RFID initiatives because of consumer backlash and many more organizations hesitant to proceed further.1 Since that time, a number of U.S.- and European-based large retailers have either adopted RFID or conducted trials.2 Whereas privacy may not be the single biggest issue stifling the deployment of RFID, it has acted to delay uptake in the retail industry.3 This paper explores whether an appropriate harmonization between consumer privacy, value, and perceived control can be established for the use of RFID in retail.

There are three vital considerations in achieving this aim: (1) how consumer awareness influences perceptions, and consequently the development of such a harmony; (2) the balance evident in other, similar, auto-ID technologies and services that have already been adopted successfully; and (3) how an appropriate harmonization between value, privacy, and control can be achieved. In fulfilling the aims of the study, the consumer value proposition for the use of RFID in retail will be explored. Consumer perceptions of RFID and associated privacy issues will also be investigated. Finally, the extent to which education and awareness affect perceptions of value, privacy, and control will be measured.

RFID is best characterized as an auto-ID technology that uses radio waves to identify objects. In the context of this study, the specific RFID technology of interest is passive tags, which are tiny transponders that can be embedded or attached to an object requiring identification. These transponders, as small as a grain of rice, do not have a power source of their own; rather, they use the energy from an incoming radio frequency signal to transmit stored data to the reader. The most important characteristic of RFID technology in relation to the tagging of consumer goods is that it does not require line-ofsight positioning, which is a requirement of bar code systems. For EPC Gen 2 UHF (electronic product code generation 2 ultra high frequency) passive tags, the read range is 3.5 meters and the write range is 2 meters, depending on the RFID system setup and the environmental conditions. It is also possible to achieve reads of up to 8 meters away using these tags. The ability for RFID tags to be read covertly is the main concern among privacy advocates.

The rest of this paper is organized as follows. In the next section, definitions of privacy, value, and control are provided in addition to a survey of related RFID works. Then, the methodology used in the current study is briefly described. In the following section, four widely adopted auto-ID applications are presented using a literature survey to explore the actual privacy, value, and control dynamics that have led to consumer acceptance of these auto-ID technologies. In the next section, the results of an online survey investigating consumer perceptions of RFID in retail are presented and a comparison is made between the qualitative and quantitative findings. In the following section, the principal outcomes of the study are discussed. A brief summary of the material presented concludes the paper.

Previous Works

The classic definition of privacy is provided by Westin, as the ‘‘claim of individuals, groups, or institutions to determine for themselves when, how, and to what extent information about them is communicated to others.’’4 This study is primarily focused on information privacy, which is described by Clarke as ‘‘the interest an individual has in controlling, or at least significantly influencing, the handling of data about themselves.’’5 Of primary concern in regard to RFID usage in retail is the collection of personal information that pertains to consumer shopping preferences, actions, and behavior. It is the collection, use, and disclosure of this information, particularly when it may be incorrect or unverified, to identify, track, and monitor individuals without their awareness or express approval, that is commonly recognized as one of the most prominent threats. It is important to understand that Clarke’s definition, along with other definitions of privacy from Altman,6 Schoeman,7 and Margulis,8 all emphasize that privacy is not separate from control; rather, it is ‘‘deeply intertwined with it.’’9

Value in this study will be viewed in terms of the benefits RFID technology affords consumers. It is how an individual prizes a certain outcome against all others.10 The value proposition to consumers for RFID usage in retail is generally phrased in terms of convenience. It is an equation of all the positive factors that interest the individual. It can include cost savings, time reductions, efficiency, personalization, safety, and security, as well as convenience and other tangible and intangible benefits. Therefore, in creating a harmony of privacy, value, and control, it is a harmonization between consumer willingness to lose some degree of privacy versus the strength of the retailer’s value proposition for using the technology.11 The value proposition can essentially be seen as a combination of benefits versus risks that consumers will evaluate in their decisions and perceptions.

Inness12 is clear that in characterizing the function of privacy in terms of control or restricted access there are ramifications for the normative value we accord privacy. For the purpose of this study, control becomes a relevant dimension of RFID acceptance, because it is only through a perceived level of control over their personal information that consumers will feel their privacy is being respected.13 The level of control that is provided either through the technology or by the service provider, whether that be perceived or real, is seen as an important element that, when combined with the value proposition, can affect consumer acceptance.

The consumer acceptance of RFID has been investigated in a number of studies. Some have proposed solutions that protect and enhance privacy and afford consumers a level of control.14–16 These solutions are typically technology-based, legislative, or regulatory in nature. Despite the different privacy solutions, a number of studies critically highlight that consumer perceptions and fear of RFID technology, brought about by a lack of understanding, remain.17,18 Thus, regardless of which privacyenhancing technologies are used, the concerns from the consumer’s perspective are the same.9,19 It is apparent from such studies that the real issue becomes one of fear or other underlying motives, that, when combined with perceptions of privacy and control, motivate a consumer’s acceptance of RFID technology. One quantitative study found that consumers felt a lack of perceived control over the technology as well as a great power distance,20 and another study found that cultural dimensions affected the way in which consumers viewed the privacy threat.21

The privacy debate has developed due to the identification and tracking capabilities inherent in the RFID technology. The argument is that if the tags were to remain active after the consumer has left the store, the technology could provide retailers and manufacturers the ability to track an individual’s movement and behavior in a clandestine manner.22 This is introduced by Roussos,23 who explains the ability of the technology to silently retrieve and record unique identifiers as an important contributing factor toward consumer uneasiness. Garfinkel et al.15 discuss seven key privacy threats that arise from the capabilities of RFID: (1) action threat; (2) association threat; (3) location threat; (4) preference threat; (5) constellation threat; (6) transaction threat; and (7) breadcrumb threat (i.e., leaving a trail of actions). Such threats have given rise to much concern by privacy advocates. In 2005, Eckfeldt24 explained that many major companies around the world had already scrapped RFID plans following consumer backlash. If it were not for the ‘‘haunting cries of privacy running afoul,’’ many more companies would have tested and launched RFID initiatives.1 This can also be seen clearly in the results of a Cap Gemini Ernst & Young consumer perception study of RFID that highlighted privacy concerns as ‘‘the most significant issue among consumers in all countries.’’25

The value proposition for RFID use in retail is an important topic that underscores consumer acceptance of RFID. What is apparent in surveying the literature is that while the benefits of RFID have been clearly defined and expressed for retailers, they have not been so clearly communicated to consumers. Eckfeldt24 makes an important assertion in discussing the value of RFID to consumers: ‘‘... the difference between successful and shunned RFID applications turns on delivery of clear, tangible value to the average consumer.’’ Furthermore he stresses that in assessing consumer benefit, organizations must consider consumers’ interests above their own; otherwise, they will produce a solution that fails to provide a positive balance between risk and reward in the eyes of the consumer. He further highlights that a tangible consumer benefit is pivotal to all these solutions. McGinity1 stresses the key value to consumers: better prices and product selection brought on by better efficiency at the back end, including reduced waste, reduced shrinkage, and improved supply chain processes. However, because the systems have not been widely implemented, assessing or promoting such benefits would appear to be speculative at best.

Balancing the economic interests of business against the privacy interests of consumers is another cornerstone in the privacy debate. Culnan and Bies11 introduce the centrist perspective, whereby corporate access to information should be balanced against the legitimate rights consumers have toward protection of their privacy. In addressing this balance, the notion of second exchange is introduced, whereby consumers make a non-monetary exchange of their personal information in return for improved service, personalization, and benefits.11 Importantly, they highlight that, for both organizations and consumers to realize the benefits, consumers must be willing to disclose their personal information and thus surrender some degree of their privacy. It is proposed, therefore, that people may be willing to accept a loss of privacy as long as there is an acceptable level of risk accompanying the benefits.

This idea of balancing interests is touched on by many authors. Eckfeldt,24 for example, emphasizes the idea of risk again in stating that successful RFID applications over-compensate for any privacy fears. He furthers the idea of risk in proposing that consumers will accept the risks if the application is worth the benefits. Langheinrich’s26 discussion on privacy claims that privacy practices and goals must be balanced with the convenience or inconvenience associated with them. In balancing the interests of consumers against organizations, the important issue that seems to dominate is the balancing of convenience and other terms of value for the consumer against the privacy incursion that is inevitable in providing such applications. It must be underscored that an underlying assumption made in this study by the authors is that privacy incursions, especially in the form of breaches in information privacy, are inevitable in the adoption of any emerging mass-market technology, and even more so if that technology happens to be wireless or mobile.

Methodology

Figure 1 Conceptual framework for the auto-ID application cases

This study used a combination of qualitative and quantitative approaches; a literature survey of four auto-ID applications, and a quantitative analysis of the data collected from an online survey. The literature survey covers the mobile phone, electronic toll collection (ETC), e-passports, and loyalty programs. The online survey analyzed the consumer value proposition for the use of RFID in retail and privacy threats relative to education and awareness. The conceptual framework for the auto-ID application cases is illustrated in Figure 1.

The conceptual framework covers the main dimensions studied in the literature search and their relationships. Harmonization can be derived from the value offering of the RFID technology, some of which is inherent to the technology itself (e.g., contactless operation), and some offered by the provider of the service using the technology (e.g., fast checkout at a supermarket). The privacy threats that the technology exposes, and the degree of control individuals have over their personal information, are also considered.

Harmonization is also affected by how widely the technology is to be used; that is, whether it is for large, high-priced items only, or for mass-market products. It has been seen that the more people use a technology (i.e., the higher the penetration rate), the less individuals question the privacy risks. The balance is also affected by the environment in which the technology is to be adopted, whether that be mandated (as in the case of e-passports), or voluntary. Finally, harmonization is also affected by societal perceptions; for example, the idea of microchips attached to common objects immediately conjures notions of Big Brother, and thus a negative perception of the technology.

Data collection for the case applications used multiple sources, including documents such as books, media reports (e.g., Factiva27), journal articles, white papers, corporate information, and marketing materials. The documents were sourced from libraries (offline), databases (e.g., IEEE Xplore28), online journals (e.g., the Journal of Theoretical and Applied Electronic Commerce Research), and media organizations (e.g., the British Broadcasting Corporation), as well as corporate, governmental, and institutional Web sites. The data collection was an iterative process, starting with a broad search strategy involving the key topics under investigation, with more targeted searches conducted thereafter.

Data collection for the online survey was administered at www.rfidsurvey.org for a period of 75 days, from July 10, 2007, through September 23, 2007. The online survey was openly accessible to all Internet users. In addition, targeted recruitment was undertaken in the form of electronic and physical mailings. The data collected in the online survey was based on 28 questions structured into four separate sections. The first section asked for general demographic information as well as information about the participants’ awareness and education. The second section queried participant perceptions of the consumer value proposition for the use of RFID in retail, asking participants to rank both awareness and importance against a list of suggested RFID benefits. The third section focused on assessing value and privacy in regard to a number of other technologies such as mobile phones, smart cards, loyalty programs, e-passports, GPS car navigation, and electronic toll collection. Four of these technologies are featured in the case study analyses. The final section of the survey asked questions about perceptions of privacy threats due to the use of RFID. Presented with a list of threats, participants were asked to rank awareness and concern of such threats. During the survey, respondents were given several opportunities to reply by way of open comments.

Qualitative content analysis was used to discover similarities between the four auto-ID application cases under investigation. Toward this end, the cases were structured in the same manner, around the themes of privacy, value, and control. The analysis focused on the significance of the technology given its penetration and usage rates, despite the presence of privacy threats, and the outcome is presented in narrative form. The text-mining tool Leximancer29 was used to analyze the documents collected, including the open comments provided by survey respondents. Leximancer assisted in uncovering the main concepts contained within the text and showed how these were interrelated.

The purpose of the statistical survey analysis was to uncover the perceptions held by participants toward RFID in retail, its potential threats, and its potential value given a number of typical usage scenarios. Perceptions of threat and value were also analyzed with regard to a number of other auto-ID technologies. Inferences were drawn on the population being studied by finding correlations using rating scales to reflect the real-world nature of the research. Given the use of the Likert approach, readers should note that the researchers were not working with quantities that provided precise measurements but working with rating scales (correlations of which provide general indications only). Using JMP* Statistical Discovery Software from SAS Institute, Inc., a common score for RFID value and threat, as well as value-and-threat scores for other auto-ID technologies, was determined by aggregating the rankings given by participants to relevant questions. The participants’ awareness of RFID and its potential use was also found in this way, using linear regression analysis.

The significance probability of the test (represented as a p-value) is a measure of how likely or unlikely it is to experience the observed data if the null hypothesis is true. The p-value is the area under the null distribution curve that is in bigger disagreement with the null hypothesis than the observed test statistic. When the p-value is less than 0.05, the result of the test is said to be statistically significant. When the p-value is less than 0.01, the result of the test is said to be highly statistically significant. The relationships between variables that were particularly significant in the data studied are illustrated using bivariate plots.

Auto-ID Applications

This section will present auto-ID application cases that explore the adoption and acceptance of a number of technologies and services within the context of privacy, value, and control.30

Mobile phone

The value proposition of the mobile phone extends from the convenience offered by its inherent mobility. In a study conducted by Hakkila and Chatfield31 regarding perceptions of mobile phone privacy, it was shown that greater than 82 percent of respondents considered their mobile phone a ‘‘private device.’’ The mobile phone presents a number of unique privacy threats, yet such privacy threats are seldom discussed or thought of by end users.32 Many citizens in the U.S., for example, are completely unaware that government authorities can track their movements by monitoring the signals that are emitted from the handset.33 The mobile phone also presents other privacy concerns in regard to the interception of signals by unauthorized persons.34 Theoretically, users can exercise control over other parties tracking their location by simply turning off their phones. However, in doing so, they prevent access to the features of the phone that provide the value in the first place.

Electronic toll collection

The key value proposition that electronic toll collection (ETC) systems offer is convenience and time saving. Such a system eliminates the burden to have cash available to make toll payments and provides individuals and corporations the convenience of an account that can provide better tracking of toll expenditure with more convenient payment options.35 Caldwell36 highlights two privacy concerns with regard to ETC. The first is the illegitimate use of drivers’ personal information related to payments, movement, and driving habits, which could become accessible if electronic records are compromised through a ‘‘cyber break-in.’’ This has been demonstrated on numerous occasions, such as the incident in which programmers were able to view ETC account details for subscribers in several countries, including one of the largest ETC systems in the United States.37 The second concern is the legitimate use of ETC account information by government authorities or private vendors that can use the information to monitor driving patterns and behavior of thousands of motorists. The concern also applies to other potential uses, such as traffic surveillance being used to detect speeding violations or stolen vehicles.38 Court cases in the U.S. have already demonstrated the potential for toll-tracking information to be used to verify an individual’s whereabouts and movements. The states of Delaware, Illinois, Indiana, Maryland, Massachusetts, New York, and Virginia have all released E-ZPass toll records in response to court orders for civil matters such as divorce. The states of Maine, New Hampshire, New Jersey, and Pennsylvania only release electronic toll records for criminal cases.39

e-passports

The greatest value of the e-passport, as stressed by most issuing authorities, is the enhanced security it is purported to provide through the digital storage of passport information.40 Certainly, given the current level of importance placed on national security, governments have been keen to introduce this technology as a means of providing more stringent monitoring of individuals entering and exiting the country.

The privacy concerns surrounding e-passports are primarily related to the ability to access passport information without contact, a capability afforded by the use of RFID to store the data contents of the passport. Juels, Molnar, and Wagner41 identify six key areas of concern: (1) clandestine scanning; (2) clandestine tracking; (3) skimming and cloning; (4) eavesdropping; (5) biometric data leakage; and (6) cryptographic weaknesses. The main issue of the e-passport is that the International Civil Aviation Authority (ICAO) does not require authentication or encryption for communications between the reader system and the e-passport. In locations where passports are frequently open, this could allow for eavesdropping. Theoretically, the unique identifier (ID) stored on the microchip could identify individuals and be used for tracking. Passports could even be cloned, because the digital signatures cannot tie the data to a particular passport. Once a reader has the key, there is no mechanism for revoking access, thus giving the reader the ability to scan the passport in perpetuity. Globally, it is reported that more than 50 million e-passports have been issued, which suggests that despite privacy concerns, the technology has undoubtedly been deployed successfully.42 Some states have mandated that the contactless microchip be shielded by a metal jacket to prevent the chip from being read when the passport is closed.43 If the shield is not provided, a sheet of aluminum foil will equally prevent unauthorized access of personal data on the e-passport.44

The media have been quick to highlight potential failures with the technology, such as the demonstration by a hacker who successfully cloned a U.S. e-passport and then dumped the contents onto an ordinary contactless smart card.45 A further threat was exposed when programmers demonstrated how an explosive device connected to an RFID reader could be triggered when a U.S. citizen carrying an e-passport came within reach of the reader.45 Given the mandatory nature of passports, there is very little individuals can do to avoid using them when traveling abroad. There is also little an individual can do to control how government authorities access and use the information on the passport when they are entering a foreign country.

Loyalty programs

In the case of loyalty programs, the value proposition is critical for encouraging consumer use and for developing the brand loyalty that the programs aim to achieve. A number of factors that determine such value in a loyalty program are described by Yi and Jeon.46 They include: (1) the cash value of rewards; (2) the choice of rewards; (3) the aspirational value of rewards; (4) the likelihood of achieving the rewards; and (5) how easy the loyalty scheme is to use.

The major privacy threat that extends from the use of loyalty programs is the ability to tie purchases of specific products to individual consumers and monitor their purchasing behavior over time. A study conducted by Graeff and Harmon47 found that in regard to loyalty programs, consumer perceptions were typically positive and most consumers did not associate such schemes with the collection and use of personal information. Loyalty programs are the ultimate demonstration of the trade-off consumers make of their privacy in order to gain something of value: benefits, rewards, convenience, or savings.48

A key element of consumer loyalty programs is their opt-in nature. Consumers are also given control over their personal information by government regulations, which in most countries grant consumers the right to know exactly what information retailers are collecting and how it is being used.

Discussion

It would appear, given the widespread use of the four auto-ID applications, that privacy has not been a barrier to their adoption and consequent acceptance by society. While the privacy concerns still exist and indeed, many individuals remain concerned about their privacy in relation to such technologies and services, on the whole it would seem that consumers have accepted each application because either the value proposition or level of control present balances against the privacy issues (mobile phones, ETC, and loyalty programs), or participation (usage) is mandatory and the appropriate safeguards to privacy are in place (e.g., e-passports).

Using Garfinkel et al.’s paradigm,15 action can be inferred by monitoring the mobile phone location, or monitoring tag usage at tollways, or monitoring passport usage, or inferred by the use of loyalty cards or the redemption of rewards. Association is prevalent in being able to identify an end user through the international mobile equipment identity (IMEI) in a mobile phone, through the tag ID or account number for tollways, through the e-passport ID number, and through the membership number on loyalty schemes. In terms of location, a mobile phone can be found through triangulation or using the Global Positioning System (GPS) chipset in the handset. The location of tags in tollways is also collected at each ETC entry and exit gantry. The location of an e-passport is established each time it is read by authorities or a reader device. For loyalty programs, the location can be established each time the card is used.

In the case of preferences, a mobile service provider has a list of features into which the user has opted. There are no preferences for ETC or e-passports. Loyalty programs allow for detailed consumer preferences to be analyzed by monitoring purchases and behavior. Information transactions are recorded by all the auto-ID applications studied. However, the loyalty card program is the only case investigated where transactions carry a value related to a monetary measure or rewards-based points scale. With respect to privacy, the breadcrumb attribute is the most invasive in terms of privacy threats. In the case of a mobile phone, a trail of actions can be inferred by the handset location or subscriber usage patterns. For ETC, a trail of actions can be generated by logging the location of the vehicle at entry and exit readers with timestamps. For the e-passport, each time it is read, the location is recorded. And for loyalty programs, a trail is automatically created of individual purchases at the point of sale. Different auto-ID applications have varying capacity to record location information, from the mobile phone that can be tracked 24-7, to the RFID in ETC that can be read several times per day on average, to the e-passport that is read at border checkpoints.

In the case of the mobile phone, the ubiquity in value terms would explain the lack of concerns consumers have toward their privacy in regard to its usage. For ETC, individuals have embraced the convenience aspects and it would seem that the ease of use of the technology (simply install the tag and forget about it) has again resulted in a general lack of concern about privacy issues. Loyalty programs are also clearly driven by their value to consumers. Of the four case studies discussed, the e-passport is the only one in which usage is almost completely mandatory for those wishing to travel internationally and also where individuals have very little control over how it is used by authorities. A summary of the key elements of value, privacy, and control for each of these technologies is provided in Table 1. For the greater part, the auto-ID technology in question provides value to the consumer by providing increased convenience. Consumers trade this value with the possibility of mobile telephone intercepts by lawful and unlawful parties, the potential to clone a tag, and the provision of personal biometric details. It is consumers’ perceived level of control of their personal information that can influence the value gained by opting in or out of a service. A key outcome that arises from the case studies presented is the varying relationship between three elements (privacy, value, and control). It is clear that in order to gain acceptance, privacy issues must be offset by value and control.

Table 1. Key elements of value, privacy, and control

In the case of mobile phones, it is evident that a somewhat low level of control is acceptable, given the relatively low vulnerability of individual privacy and the medium level of value the technology provides. With ETC, the vulnerability of user privacy is considered to be in the medium range, yet as users can exercise some degree of control over their privacy by removing the tag or opting to use alternative routes or payment methods, control is also depicted as being in the medium range. This medium range in regard to privacy and control is offset by a high level of value evident in the convenience the technology affords. With regard to e-passports, the government provides very little control. Furthermore, the value offered to the individual is, in real terms, also very low. Finally, with loyalty programs, a high vulnerability of individual privacy that arises from the vast amount of personal information collected is offset by a high level of control offered by providers by allowing consumers to freely opt out of such programs. The privacy risk is also further offset by the high level of value that such schemes must offer to encourage consumers to participate.

In the case of mobile phones, ETC, and loyalty programs, it is apparent that acceptance had to be earned through a favorable balance that was offered to consumers. In the case of e-passports, where the balance is unfavorable, acceptance was not generally required, as the technology was made mandatory by government authorities and the ICAO.

 

Analysis of Online Survey Data

The threats listed in the survey are potential threats of RFID (i.e., perceived threats) that have been drawn out from the literature as the major causes for consumer concern over the use of RFID in retail. Awareness refers to the aggregated score of each survey participant’s responses to a number of questions that dealt with perceptions of RFID and other auto-ID technologies. Specifically, the awareness score was calculated by the sum of responses in which participants ranked, using a Likert49 scale of 1 to 5, their knowledge on a list of 12 RFID related topics.

Sample respondents

Figure 2 Relationship between age and consumer awareness of RFID in retail

There were 142 survey responses in the pilot study. The majority (61.1 percent) of surveys were completed by Australians. The U.S. had the second largest number of responses (27.4 percent), with other responses recorded from countries such as Canada, Germany, Spain, and the United Arab Emirates.

Figure 2 aims to demonstrate the role that age plays in determining the level of awareness toward RFID. In analyzing the relationship between age and awareness, there is a highly significant relationship (p ¼ 0.0008) between a respondent’s age and his or her associated level of awareness. The data shows that awareness decreases with age, which is to be expected given that younger respondents are more likely to have been exposed to the technology, or have a heightened awareness of the possibilities and issues such technology represents.

Figure 3 Relationship between consumer awareness and value proposition for RFID in retail

Figure 3 Relationship between consumer awareness and value proposition for RFID in retail

Figure 3 shows the relationship between awareness and the consumer value proposition for RFID as being statistically significant (p ¼ 0.0337). It is seen that as awareness increases, the participants’ rankings of RFID value decreases. This relationship suggests that those individuals who are highly aware of the technology are less likely to embrace the value of technology, as they are at the same time balancing the value against their perception of the privacy threats of the technology. Individuals who are less aware of the technology are more easily swayed by the value the technology provides.

Surprisingly, it would seem that awareness plays little role in an individual’s ability to perceive the privacy threats that the technology could introduce if it were to be implemented. This suggests perhaps that participants, regardless of their awareness of RFID, are able to appreciate the privacy issues based on their previous life experiences, particularly with other technologies presenting similar issues.

Figure 4 Consumer concern over privacy: RFID in retail versus other auto-ID applications

The results also indicate that there is some statistical significance in the relationship between RFID value and privacy threat. The higher an individual ranks the potential value of RFID, the lower they rank the potential privacy threat. It would suggest that elements of consumer value proposition for RFID, such as convenience, may override any potential privacy threats. Thus, presenting a clear value for RFID could be seen as important in countering any potential losses in privacy.

A key element of the survey was the ranking participants provided on both value and privacy concerns in regard to a number of other related technologies that have enjoyed widespread adoption (Figure 4). There was a highly significant relationship (p ¼ 0.0028) found between the perceived privacy threat of these other technologies and RFID usage in retail. In essence, respondents who were concerned about their privacy in relation to the other technologies were just as likely to be concerned about their privacy if RFID were to be adopted in retail.

Analysis of open comments

Analysis of the comments revealed a great range of attitudes, ranging from individuals who were strongly focused on potential privacy issues, to individuals who saw the technology as something quite positive and thus balanced this against the potential privacy issues. There were also many individuals who proposed safeguards that would need to be in place to make the technology acceptable.

In regard to privacy, there were a number of respondents who voiced their concerns. Comments such as, ‘‘I should have my right to privacy,’’ ‘‘... it invades on our personal freedoms,’’ ‘‘It’s too obtrusive,’’ and ‘‘... this technology is a violation of people’s right to privacy’’ clearly express strong feelings toward the potential of RFID to erode privacy of the individual. Many individuals also stressed that while they could see the value, or see the positives, they were not convinced that potential privacy issues would be managed effectively. This is well represented in the comment, ‘‘the benefits ascribed to RFID technology for the retail trade are commendable, but I have zero confidence that they will be achieved, and, instead, consumers will be subjected to more advertising, intrusion, and loss of privacy than ever.’’

Contrarily, there were a number of respondents who clearly valued the technology despite any potential privacy issues. This is illustrated by the comments, ‘‘... only someone trying to hide something or [run] from something would think this system is not a positive thing,’’ ‘‘... the benefits for consumers ... far outweigh the privacy issues that are envisaged,’’ and ‘‘... the privacy issues would sort themselves out in time.’’

A few respondents critically pointed out that indeed, this study assumed RFID technology would replace the bar code at some point. They also stated that the technologies were more complementary to each other, and that the value of placing RFID tags on every item is not justified by the present cost in doing so.

It would seem that the majority of users approach the technology with the idea that control would best balance the value against the privacy issues. The clear majority of comments expressed that the design of RFID systems should incorporate privacy protection from the outset. A common theme is seen in the comment, ‘‘if proper privacy and security architectures were implemented and enforced, the deployment of RFID systems need not be so problematic ... ’’ And again from another respondent, ‘‘if privacy concerns were taken into account and proper privacy-enhancing technologies were implemented and used, we could have the benefits without the drawbacks ... ’’

Regulation and legislation were also pointed out by a number of respondents as important means of providing individuals with control over their privacy. Some consumers noted they would be happy with using the technology provided that ‘‘the technology was adequately regulated... .’’

On the whole, it is apparent that most users are more concerned about the misuse of their information than the actual collection of it. While privacy could be protected by a range of controls, the potential for the technology (as with any technology) to be misused and abused by ‘‘the low integrity sector of society’’ represents the greatest fear.

Figure 5 Overall respondent feelings toward RFID in retail

Figure 5 Overall respondent feelings toward RFID in retail

Together with the open comments, survey participants were also asked to provide a general ranking of RFID technology as it would be used in retail. Surprisingly, given the comments made and also the fact that the mean ranking in regard to privacy threats and RFID was 77 percent, the majority of individuals were neutral to very positive toward the technology (Figure 5). It would seem that most individuals can appreciate the technology, and although the privacy issues exist, they feel that the issues can be overcome, offset, or controlled in some manner.

A number of important outcomes are evident from the statistical analysis presented in this paper. These are summarized below:

  • As awareness of RFID and its associated issues increases, the relative importance of a consumer value proposition for RFID decreases.
  • Awareness of RFID and associated issues does not affect the perception of threat due to RFID.
  • The perceived privacy threat, and value, of RFID in retail is relative to an individual’s feelings toward other technologies and services with issues similar to RFID.

The most important observation in analyzing the results from the survey is the seemingly contradictory responses provided by the respondents. It was not uncommon to find participants who identified RFID as privacy-threatening, yet also stated that they were members of a loyalty program, or that they were mobile phone users.

Survey Results

In comparing the statistical results for the auto-ID application cases, it is evident that concern surrounding the privacy threat due to RFID in retail is considerably greater than the concern participants express for other applications. Where users have little to no concern regarding privacy and technologies, as is the case with the mobile phone and ETC applications and services such as loyalty programs, concern about RFID privacy threats is higher than should be expected. The key outcome that this exposes is the lack of harmonization in the current privacy, value, and control offering that RFID in retail presents.

In the application cases discussed, it was emphasized that appropriate harmonization between value and control could offset privacy issues. This is reflected in the relatively low level of concern participants in this survey placed on such technologies and services. Thus, the high rankings of privacy threats due to RFID in retail demonstrate that more education would be required to convince consumers of the value offered and the control they could exert over RFID usage. It is, however, important to understand that these rankings were given for auto-ID applications that are already widely adopted, whereby individuals have had time to understand and experience them in the context of their own lives. The privacy threat rankings individuals gave RFID, in many cases, show the lack of awareness of RFID. If consumers were actually to experience RFID usage in retail and place it in context with their own activities, it could be seen that rankings of the privacy threats may be significantly different, and perhaps more in line with the other auto-ID applications highlighted.

Therefore, it could be concluded, based on all the key results presented in this paper, that creating a favorable harmony of privacy, value, and control is perhaps an unrealistic notion when the technology has yet to be deployed. When there is such a divergent level of awareness among the greater population, striking a balance that is acceptable to all is an improbable task. It is therefore suggested that acceptance of RFID in retail may ultimately come over time, after adoption, as users become intimately experienced with its usage, or observe other user experiences. Consequently, privacy, value, and control are adjustable measures based on the feedback and behaviors of society in a given context and specific point in time. In that sense, harmonization will eventually occur with RFID in retail, just as it was shown with the auto-ID application cases presented.

The principal outcomes of this study can be summarized as follows:

  • The value proposition for RFID has not been well communicated to consumers.
  • Concerns surrounding RFID in retail were disproportionately higher than other previously adopted auto-ID applications despite similar privacy issues.
  • A harmonization between privacy, value, and control is unrealistic prior to adoption and can only be achieved once consumers can be educated through experience with the technology.

The preliminary findings of this study suggest that the harmonization between privacy, value, and control is largely dependent on individuals and their background (e.g., age), the type of technology being deployed (i.e., level of perceived invasiveness), and the type of provider (i.e., government or commercial entity). The results indicate that the perceived value and privacy threats posed by RFID in retail are commensurate with an individual’s pre-existing feelings toward other, similar, technologies. As was shown, privacy-related issues per se have not been a barrier to widespread adoption of auto-ID applications. On this point, the level of consumer awareness of RFID in retail does not seem to affect perceptions of privacy threats. It does, however, affect perceptions of value. Thus, a favorable harmonization whereby privacy is offset by value and control has been shown to encourage consumer acceptance.

The auto-ID application cases highlighted the importance of a harmonization between privacy, value, and control in influencing consumer acceptance and adoption. The online survey demonstrated the effect awareness has on perceptions and the disproportionately high rankings given for RFID privacy concerns.

The most significant outcome drawn from the combined analysis of the cases and the online survey is that achieving a harmony of privacy, value, and control for RFID adoption in retail is unrealistic at this point in time. With such differing levels of awareness and education, differing expectations, and differing perceptions, achieving a harmony that is favorable to all consumers now would be an improbable task. It is also evident in reviewing the literature that there have already been significant attempts to address privacy issues and provide individuals with a degree of control, yet the privacy concern still remains. This furthers the notion that it is unlikely that privacy concerns can be resolved prior to the technology’s adoption and use by consumers.

Figure 6 Harmonizing value, privacy, and control through the adoption process

RFID in retail can certainly achieve a favorable harmonization, one that offsets privacy risks with significant value and consumer control. It is more realistic, however, for this harmony to be achieved after adoption, when consumers can be educated through their experiences, and whereby society will consequently shape the balance as the impact of the technology becomes more evident. Figure 6 illustrates that to achieve harmonization there must first be a strong value proposition driving adoption in the first place.

Conclusion

In a society where it seems we are increasingly surrounded by technologies, governments, and institutions monitoring every move we make and collecting vast amounts of personal information, privacy has grown to become an ardently debated topic. Each individual living within a civil society has a right to privacy, yet in the wake of technologies that afford us great value, there will always be some loss of privacy. This study has not sought to dismiss privacy concerns, or argue to protect privacy, but rather to address it in the realistic context it plays in an environment of technological innovation driven by society itself. Ultimately, acceptance of a technology with privacy issues will always be a balancing act, a harmonization of privacy, value, and control.

Cited References and Notes

1. M. McGinity, ‘‘RFID: Is This Game of Tag Fair Play?’’ Communications of the ACM 47, 15–18 (2004).

2. J. Whitaker, S. Mithas, and M. S. Krishnan, ‘‘A Field Study of RFID Deployment and Return Expectations,’’ Production and Operations Management 16, No. 5, 599–612 (2007).

3. K. Michael and L. McCathie, ‘‘The Pros and Cons of RFID in Supply Chain Management,’’ Proceedings of the 2005 International Conference on Mobile Business (ICMB 2005), July 11–13, 2005, Sydney, IEEE Computer Society (2005) pp. 623–629.

4. A. F. Westin, Privacy and Freedom, The Bodley Head Ltd. (1970).

5. R. Clarke, ‘‘Information Technology and Dataveillance,’’ Communications of the ACM 31, No. 5, 498–512 (1998).

6. I. Altman, The Environment and Social Behavior: Privacy, Personal Space, Territory, Crowding, Brooks/Cole Publishing, Monterey, CA (1975).

7. F. D. Schoeman, Philosophical Dimensions of Privacy: An Anthology, Cambridge University Press, Cambridge, UK (1984).

8. S. T. Margulis, Contemporary Perspectives on Privacy: Social, Psychological, Political, Blackwell Publishing, London (2003).

9. S. Spiekermann, ‘‘Perceived Control: Scales for Privacy in Ubiquitous Computing Environments,’’ Proceedings of the 10th International Conference on User Modeling, Edinburgh, Scotland (2005).

10. B. D. Renegar and K. Michael, ‘‘The RFID Value Proposition,’’ Proceedings of the Sixth CollECTeR Iberoamerica: Collaborative Electronic Communications and eCommerce Technology Research, June 25–27, 2008, Madrid (2008) pp. 1–10.

11. M. J. Culnan and R. J. Bies, ‘‘Consumer Privacy: Balancing Economic and Justice Considerations,’’ Journal of Social Issues 59, No. 2, 323–342 (2003).

12. J. C. Inness, Privacy, Intimacy and Isolation, Oxford University Press, New York (1996).

13. J. R. Averill, ‘‘Personal Control over Aversive Stimuli and its Relationship to Stress,’’ Psychological Bulletin 80, No. 4, 286–303 (1973).

14. R. Bansal, ‘‘Now You See It and Now You Don’t [RFID Technology],’’ IEEE Microwave Magazine 5, No. 4, 32–34 (2004).

15. S. L. Garfinkel, A. Juels, and R. Pappu, ‘‘RFID Privacy: An Overview of Problems and Proposed Solutions,’’ IEEE Security & Privacy 3, No. 3, 34–43 (2005).

16. L. Hyangjin and K. Jeeyeon, ‘‘Privacy Threats and Issues in Mobile RFID,’’ Proceedings of the First International Conference on Availability, Reliability and Security (ARES 2006), April 20–22 2006, Vienna, IEEE Computer Society (2006), pp. 510–514.

17. G. Roussos and T. Moussouri, ‘‘Consumer Perceptions of Privacy, Security and Trust in Ubiquitous Commerce,’’ Personal and Ubiquitous Computing 8, No. 6, 416–429 (2004).

18. O. Gunther and S. Spiekermann, ‘‘RFID and the Perception of Control: The Consumer’s View,’’ Communications of the ACM 48, No. 9, 73–76 (2005).

19. S. Spiekermann, User Control in Ubiquitous Computing: Design Alternatives and User Acceptance, Shaker Verlag, Aachen, Germany (2008).

20. G. Ng-Kruelle, P. A. Swatman, D. S. Rebne, and J. F. Hampe, ‘‘The Price of Convenience: Privacy and Mobile Commerce,’’ Quarterly Journal of Electronic Commerce 3, No. 3, 273–385 (2002).

21. G. Ng-Kruelle, P. A. Swatman, J. F. Hampe, and D. S. Rebne, ‘‘Biometrics and e-Identity (e-passport) in the European Union: End-user Perspectives on the Adoption of a Controversial Innovation,’’ Journal of Theoretical and Applied Electronic Commerce Research 1, No. 2, 12–35 (2006).

22. B. J. Alfonsi, ‘‘Privacy Debate Centers on Radio Frequency Identification,’’ IEEE Security & Privacy 2, No. 2, 12 (2004).

23. G. Roussos, ‘‘Enabling RFID in Retail,’’ Computer 39, No. 3, 25–30 (2006).

24. B. Eckfeldt, ‘‘What Does RFID Do for the Consumer?’’ Communications of the ACM 48, No. 9, 77–79 (2005).

25. C. Perakslis and R. Wolk, ‘‘Social Acceptance of RFID as a Biometric Security Method,’’ IEEE Technology and Society Magazine 25, No. 3, 34–42 (2006).

26. M. Langheinrich, ‘‘Privacy by Design: Principles of Privacy-Aware Ubiquitous Systems,’’ Proceedings of the 3rd International Conference on Ubiquitous Computing, (Ubicomp 2001), Atlanta, September 30–October 2, 2001, Lecture Notes in Computer Science 2201, Springer (2001), pp. 273–291.

27. Factiva, Dow Jones & Co., http://www.factiva.com/.

28. IEEE Xplore, IEEE, http://ieeexplore.ieee.org/Xplore/ guesthome.jsp.

29. Leximancer (2008), http://www.leximancer.com/.

30. B. D. Renegar, K. Michael, and M. G. Michael, ‘‘Privacy, Value and Control Issues in Four Mobile Business Applications,’’ Proceedings of the Seventh International Conference on Mobile Business, July 7–8, 2008, Barcelona (2008), pp. 30–40.

31. J. Hakkila and C. Chatfield, ‘‘ ‘It’s Like If You Opened Someone Else’s Letter’: User Perceived Privacy and Social Practices with SMS Communication,’’ Proceedings of the 7th International Conference on Human Computer Interaction with Mobile Devices & Services, September 19–22, 2005, Salzburg, Austria, ACM, New York (2005), pp. 219–222.

32. N. Swartz, ‘‘Mobile Phone Tracking Scrutinized,’’ Information Management Journal 40, No. 16 (2006), http:// www.entrepreneur.com/tradejournals/article/ 184698661.html.

33. W. A. Herbert, ‘‘No Direction Home: Will the Law Keep Pace with Human Tracking Technology to Protect Individual Privacy and Stop Geoslavery?’’ I/S: a Journal of Law and Policy for the Information Society 2, No. 2, 409–473 (2007).

34. R. Whitaker, The End of Privacy: How Total Surveillance Is Becoming a Reality, The New Press, New York (1999).

35. P. Hills and P. Blythe, ‘‘Paying Your Way [Road Tolls],’’ The IEE Review 35, No. 10, 377–381 (1989).

36. C. Caldwell, ‘‘A Pass on Privacy?’’ The New York Times, July 17, 2005, http://www.nytimes.com/2005/07/17/ magazine/17WWLN.html?ex¼1279339200&en¼ c1f10d3de06adea6&ei¼5088.

37. A. McCluskey, ‘‘Position Paper: Business Ethics,’’ BT Financial Group 3, 1–5 (2004).

38. IBI Group, ‘‘Background Paper #8: Toll Technology Considerations, Opportunities, and Risks,’’ Washington State Comprehensive Tolling Study: Final Report 2, 1–33 (2006).

39. V. D. Hunt, A. Puglia, and M. Puglia, ‘‘RFID Technology in Homeland Security, Law Enforcement, and Corrections,’’ in V. D. Hunt, RFID: A Guide to Radio Frequency Identification, Technology Research Corp., New York (2007), pp. 67–82.

40. M. Meingast, J. King, and D. K. Mulligan, ‘‘Embedded RFID and Everyday Things: A Case Study of the Security and Privacy Risks of the U.S. e-Passport,’’ IEEE International Conference on RFID, March 26–28, Grapevine, Texas (2007), pp. 7–14.

41. A. Juels, D. Molnar, and D. Wagner, ‘‘Security and Privacy Issues in e-Passports,’’ First International Conference on Security and Privacy for Emerging Areas in Communication Networks, Athens, Greece (2005), pp. 74–88.

42. C. Edwards, ‘‘Borderlands of Confusion [Biometric Passports],’’ The IEE Review 51, No. 11, 34–37 (2005).

43. ‘‘Machine Readable Travel Documents (MRTDs): History, Interoperability, and Implementation,’’ working paper, International Civil Aviation Organization, Montreal (March 23, 2007), http://www.icao.int/icao/en/atb/ sgm/mrtd/TAG_MRTD17/TagMrtd17_WP016.pdf.

44. M. Sirotich, ‘‘ePassport Security Under the Microscope,’’ The Second Workshop on the Social Implications of National Security: From Dataveillance to Uberveillance and the Realpolitik of the Transparent Society 2, K. Michael and M. G. Michael, Eds., University of Wollongong, Wollongong, Australia (2007), pp. 257–280.

45. K. Zetter, ‘‘Hackers Clone E-Passports,’’ Wired News, August 3, 2006, http://www.wired.com/science/ discoveries/news/2006/08/71521?currentPage¼1.

46. Y. Yi and H. Jeon, ‘‘Effects of Loyalty Programs on Value Perception, Program Loyalty, and Brand Loyalty,’’ Journal of the Academy of Marketing Science 31, No. 3, 229–240 (2003).

47. T. R. Graeff and S. Harmon, ‘‘Collecting and Using Personal Data: Consumers’ Awareness and Concerns,’’ Journal of Consumer Marketing 19, No. 4/5, 302–318 (2002).

48. D. H. Nguyen, A. Kobsa, and G. R. Hayes, ‘‘An Empirical Investigation of Concerns of Everyday Tracking and Recording Technologies,’’ Proceedings of the 10th International Conference on Ubiquitous Computing 344, Seoul, Korea, ACM, New York (2008), pp. 182–191.

49. Attitude measurement used in surveys in which, in response to questions, respondents select from a set of typically five values, such as from complete agreement to complete disagreement, with no opinion in the middle.

Benjamin D. Renegar IBM Global Business Services, IBM Centre, 601 Pacific Highway, St. Leonards, NSW, Australia 2065 (brenegar@au1.ibm.com). Ben Renegar is a recent graduate from the University of Wollongong, having completed a Bachelor of Information and Communication Technology degree at the end of 2007 with the award of first-class honors. For this degree program, he completed a thesis on RFID adoption in the retail industry with a focus on the harmonization of value, privacy, and control. He was also awarded the PriceWaterhouseCoopers award for the highest grade in this program. He was employed by IBM as a Graduate Consultant in the Application Innovation Service Delivery organization in 2008.

Katina Michael University of Wollongong, NSW, Australia 2500 (katina@uow.edu.au). Dr. Michael is a Senior Lecturer in the School of Information Systems and Technology in the Faculty of Informatics at the University of Wollongong. She received a Bachelor of Information Technology degree from the University of Technology, Sydney (UTS) in 1996 and a Ph.D. degree in information technology and communications from the University of Wollongong in 2003. Before joining the University of Wollongong in 2002 to teach and conduct research in e-Business, she worked as a senior network and business planner at Nortel Networks. In 2000, Katina received the Nortel top talent award for work completed on 3G mobile networks in Asia. She is a senior member of the IEEE and a Board Member of the Australian Privacy Foundation.

The Auto-ID Trajectory - Chapter Ten: Conclusion

The principal conclusions from the findings given in chapter nine are threefold. First, that an evolutionary process of development is present in the auto-ID technology system (TS). Incremental steps either by way of technological recombinations or mutations have lead to revolutionary changes in the auto-ID industry- both at the device level and at the application level. The evolutionary process in the auto-ID TS does not imply a ‘survival of the fittest’ approach,[1] rather a model of coexistence where each particular auto-ID technique has a path which ultimately influences the success of the whole industry. The patterns of migration, integration and convergence can be considered either mutations or recombinations of existing auto-ID techniques for the creation of new auto-ID innovations. Second, that forecasting technological innovations is important in predicting future trends based on past and current events. Analysing the process of innovation between intervals of widespread diffusion of individual auto-ID technologies sheds light on the auto-ID trajectory. Third, that technology is autonomous by nature has been shown by the changes in uses of auto-ID; from non-living to living things, from government to commercial applications, and from external identification devices in the form of tags and badges to medical implants inserted under the skin.

Read More