IEEE 2089 Age Appropriate Digital Services for Children

Abstract

Online digital services have changed the way that people interact. Companies provide apps for download allowing users of any age to experience them through smartphones and tablets among other devices. To date, company policies have acted as pseudo-guidelines for recommended use. But what happens when apps that were never designed for children are acquired and used by them? To mitigate potential risks the IEEE 2089–2021 standard was developed- an age appropriate digital services framework for children. The standard stipulates the need for a risk-based age appropriate register by which developers can do away with potential intolerable harms on children during the design phase, and keep track of unintended hazards, in order to counteract ongoing negative impacts on children, allowing them to thrive and flourish. Supplementing international law, state regulations, and company policies related to acceptable use, IEEE 2089–2021 provides a benchmark for how children’s apps should be designed based on the 5Rights Principles. Technical standards can be considered a type of soft law, supplementing hard law like treaties or acts, and even non-legally binding instruments like declarations and policies. Together this panoply of safeguards can mitigate the potential for flaws in product development, ranging from data privacy breaches, location tracking default features, nudging toward in-gaming purchases and auto scrolling, child labor toward data annotation, and adverse metaverse experiences. But given the rapidity of product development cycles, it is technical standards that can have the most immediate effect on the pacing problem ensuring that child rights impact assessments (CRIA) are implemented in practice.

Biographies

Prof. Dr. Katina Michael (Senior Member, IEEE) is the inaugural program director of the MBA (Technology and Digital Strategy) at The University of Sydney Business School. She is professor of Strategy, Innovation and Technology. She connects technical, policy, and public audiences, raising awareness of sociotechnical challenges and how to address them through human-centered design. She was the IEEE 2089–2021 Working Group Chair for the Age Appropriate Digital Services Framework Based on the 5Rights Principles for Children from 2019 to 2021. She was also involved in the development of the European Reference Document for Children’s Protection and Well-being Online. The CWA published by CEN/CENELEC localized the IEEE 2089 standard within the European context, where it addressed specific needs identified in various EU policies and regulations.

Assoc. Prof. Dr. Thao Ha is a developmental psychologist and director of the @HEART Lab (Healthy Experiences Across Relationships and Transitions Lab) at Arizona State University. Her NIH- and foundation-funded research investigates how emerging technologies, including social media, AI, and VR, are impacting adolescent romantic relationships and mental health during a developmentally vulnerable period. By examining how teens navigate love, breakups, and relationship dynamics within their broader social worlds (including peers, parents, and educators), Dr. Ha identifies key risk and resilience factors that influence long-term mental health. She conducts culturally grounded, community-engaged research with youth and families, providing critical insights for designing technology, policies, and interventions that foster emotionally healthy relationships in both digital and physical spaces.

Dr. Taren McGray is a behavioral and social health scientist and the post-doctoral researcher in the @HEART lab (Healthy Experiences Across Relationships and Transitions Lab) at Arizona State University. Her research examines the epidemiology and measurement of interpersonal violence in underrepresented global communities, with a particular emphasis on emerging technology-facilitated forms of violence. Grounded in community-engaged, socio-ecological, and intersectional frameworks, Dr. McGray’s research serves to translate the lived experiences of survivor-victims and at-risk individuals into actionable advocacy, policy, and prevention efforts to foster healthy and safe digital spaces, reduce digital and in-person violence, and improve mental health outcomes.

Ms Ruth Lewis: Director of Technology Foresight Consulting, Bachelor of Engineering (Electrical), Graduate Diploma Digital Communications, Master of Strategic Foresight, National Engineering Register. Ruth is Vice President (Standards) and the Chair of the IEEE Society on Social Implications of Technology (SSIT) Standards Committee.

Presentation Transcript

Welcome and framing

Katina Michael: Hello everyone, and welcome to today's IEEE SSIT education webinar. My name is Katina Michael. I'm a Professor of Strategy, Innovation and Technology at the University of Sydney Business School, and I was also the Chair of the IEEE 2089 Working Group.

It gives me great pleasure to deliver this webinar with Thao Ha, Taran McGrath and Ruth Lewis. But first, a special thank you to the organisers, Kalyan Sen and Dr [Krish], and to the team at large behind the education section of IEEE SSIT.

It gives me great pleasure to introduce Associate Professor Thao Ha of Arizona State University. Thao is the Director of the HEART Lab and has a background in developmental psychology. We're also honoured to have Dr Taran McGrath, a postdoctoral researcher from Arizona State University, also from the HEART Lab at ASU, and a behavioural and social health scientist. And last but not least, Ruth Lewis, Vice President of Standards for the IEEE Society on Social Implications of Technology and also the Chair of the SSIT Standards Committee. You'll learn more about our speakers as they present today, but just a note to say that Ruth Lewis is a director of technology foresight consulting here in Australia, working internationally.

Over the next hour we'll explore the social implications of innovation, the regulatory context, and how standards, regulations and policies can be applied in practice to support young children.

But first, three propositions.

Let's say that the diffusion of innovations produces social implications within communities. Let's also propose that the social implications of innovation diffusion may be either beneficial or harmful to users, end users and usees in the community. Our third proposition is that the diffusion of innovations disproportionately affects vulnerable populations, such as children, thereby necessitating additional protections. And pretty much, that's what our panel will be about today.

Types of innovation

To understand why children are so impacted, we need to consider how innovation unfolds. Henderson and Clark, dating back to 1990, identify four types of innovation: radical, modular, architectural and incremental.

To pause on each for a moment, let's think about radical innovations, such as immersive virtual reality, which can fundamentally reshape children's experiences. They put on wearable computing and they see, perhaps, in 3D.

Modular innovations are like algorithmic upgrades: they subtly alter engagement patterns. For example, if I increase the battery power or the CPU speed, I might be able to stay on a device for a longer period of time.

Architectural innovations change how systems are connected — for example, across platforms or ecosystems. Let's say a child can use a single sign-on and log in to Facebook games and other games on other platforms without having to enter a different password.

And incremental innovations are small, frequent and often invisible, yet their cumulative effect can be profound. These are the things we sometimes miss that come through in application upgrades.

Importantly, these innovations are rarely designed with children specifically in mind. Even when children are among the primary users, children are seldom asked to be part of the design process.

Children as a vulnerable population

Children are widely recognised as a vulnerable population, like the ageing population. Their cognitive, emotional and social development is ongoing, which affects how they perceive consent and the consequences that follow. Design choices that may seem trivial to adults — such as default settings, persuasive interfaces or data collection practices — can have significant effects on children, because they do not understand them. This vulnerability places a responsibility on organisations to anticipate harm rather than respond only after harm occurs. Often, when harm occurs, it is too late.

At an international level we do have children's rights articulated through instruments like the UN Convention on the Rights of the Child. This is hard law, and several articles might come to mind — UNICEF has written widely about these. Article 3, for example, is the best interests of the child. Article 12 affirms children's right to be heard. Article 16 establishes the right to privacy. Article 17 promotes children's access to reliable and appropriate information.

But General Comments on the UN Convention on the Rights of the Child are not hard law. They're not enforceable; in fact they're called soft law. These rights apply equally in digital environments. Even though the Convention predates the modern internet, these things don't change. Child rights have always remained the same; they perhaps just look different when manifested in digital services.

Hard law, soft law and the pacing problem

If we think about other soft law examples, we can think about the Universal Declaration of Human Rights. It's a declaration, not a convention, and so we call that soft law. Above and beyond soft law such as declarations or General Comments on conventions, we then have things like CRIAs — child rights impact assessments — as part of a design process, or perhaps a standard like IEEE 2089.

Other soft law examples include statements and principles, codes of conduct, codes of practice, recommendations and guidelines, policy declarations and communications, and standards and action plans. These are not enforceable, I reiterate. What is enforceable are regulations and laws. Some of these predate the internet and some have come afterwards — for example, the Age Appropriate Design Code in the UK, released in 2020, and the Digital Services Act in the European Union in 2023. But if we go through this slide, we can see some things that do predate the internet, dating back even to 1972. These are just some of the regulations and laws that may be incorporated into modern-day practices.

As we can see, technology is fast-moving, and because of this, soft law often fills the gaps left by slower legislative processes. It takes time to institute new laws, regulations and policies.

Unfortunately, companies can fail or fall short in abiding by these laws and regulations, and in the context of hard law, penalties often ensue. Some of these penalties are half a billion dollars; some may be between $20 and $40 million. We can see this in the context of child rights. For further information, please feel free to download a paper I have written, and many others that people have placed in IEEE Technology and Society Magazine and in IEEE Transactions on Technology and Society.

The question remains, however, why some organisations reoffend — and we can see this in a panoply of cases. But what's really going on here?

Despite increasing regulation, we continue to see violations affecting children, including misuse of personal data, manipulative design, and exposure to harmful content. This reflects what scholars call the pacing problem — including the scholar who coined the term, Gary Marchant, at Arizona State University's law school. This is when technological innovation advances faster than governance mechanisms can respond, and as a result organisations often face uncertainty about their responsibilities while children bear the risks. Sometimes it is unclear to organisations how far they can go.

Standards can help address this imbalance. Maybe they're not laws and regulations, maybe they're not enforceable, but they at least give a company, an organisation or a society some way of embedding foresight into the design and development process.

We can see here a paper by the Spanish author Marta Beltrán, showcased in the September 2025 Transactions on Technology and Society, looking at classifying and identifying addictive patterns in digital products — implicitly addictive persuasive design principles placed into products. When these kinds of addictive patterns are placed into children's digital services, we have issues that are not easy to solve.

For some governments around the world, we're a bit late to the table. The knee-jerk reaction, perhaps, is to ban all services, and thus there is an exclusionary principle occurring, perhaps against children, to try to curb these risks for those who are developing and under the age of 18. Youth are impacted directly when bans like the social media bans in Australia are enacted. The question, and the jury is still out, is whether this is the most effective way to go about things.

However, we have demonstrable cases, such as those recently heard — lawsuits against AI companies around the world where chatbots encouraged people to take their own life, such as the young man I'm sure we'll hear more about.

I want to spend a moment introducing Associate Professor Thao Ha and Dr Taran McGrath of the HEART Lab at Arizona State University. They have received an NIH-funded research grant for work on how emerging technologies, including social media, AI and VR, are impacting adolescent romantic relationships and mental health during a developmentally vulnerable period. I'll hand over to Dr McGrath and Associate Professor Ha.

Adolescent development and technology (HEART Lab)

Taran McGrath: Thank you so much, Katina. I'm Dr Taran McGrath and I'm here with Dr Ha. I'm the postdoctoral researcher at Arizona State University's HEART Lab and Dr Ha is the Director of the HEART Lab. HEART stands for Healthy Experiences Across Relationships and [Transitions], and that's exactly what we study. We look at adolescent relationships in particular — how teens are using technology in their relationships, and how this technology use in relationships is then associated with their mental health.

Today we're learning about this exciting new standard, and I'm just going to give you a little bit of context from our work in the HEART Lab about why being age-appropriate is important when we think about teen development.

Why adolescence matters

I want to start by talking about why we study adolescence in our lab. First and foremost, adolescence is so much more than just an age range. It represents a sensitive developmental period that is unique from both childhood and adulthood. Teens are in a highly transformational time in their life. They're beginning to explore and experience new relationships, relying on peer relationships outside of their family of origin as they did in childhood. They're starting to debut dating behaviours and generally becoming more independent in their relationships.

Adolescence is also more than a social transformation. Teen brains themselves are rapidly rewiring and growing and taking in all of these new experiences, which uniquely prepares teens to crave and explore these relationships — to learn what is normative, what's valued in their social and cultural systems, how to interact with and understand other people, and generally how to move about the world.

All of the social learning and neural growth that happens during adolescence helps them build key social-emotional competencies: a sense of identity, self-esteem, autonomy, conflict management skills, perspective-taking skills, emotional regulation and so much more. The list goes on. And these competencies, you may have guessed, are not just influential during adolescence; they're influential well into adulthood. These are things we all use, or hopefully all use — the ability to manage conflict, to understand perspectives different from our own, to regulate our emotions. These are skills that are built during adolescence and that have short- and long-term impacts on mental health and on the trajectory of future relationships.

This sets the stage for adolescence being a time when the interactions that teens are having are critically influential throughout life. But with emerging technologies, these relationships don't look the way they did 20, 10, even five years ago — even last year. We're seeing new trends emerging.

How technology reshapes teen relationships

Technology has really transformed how teens talk to each other, and in turn it's affecting how these technology-facilitated relationships are associated with mental health. There's a level of constant digital connectivity that is so normalised today: text messages, DMs, posting, liking and commenting on social media, sharing locations, and even more recently interacting with AI chatbot companions.

Globally, four in five teens have access to the internet; even in low-income countries, this is almost half of teens. In the US, the average age at which teens get their first smartphone is eight and a half years old, which means that by the time they get to adolescence, this critical phase, they not only have smartphones, they're also very experienced in how to use them. With chatbots in the US, we see one in three teens using chatbots — a lot of them on a daily basis — and one in five teens are in romantic relationships with AI. [inaudible] It's clear to see how integrated this is in their social lives.

We also want to look at how teens are using technology. Digital platforms have certain affordances, which essentially refers to the properties and design of digital platforms that change the way teens use them to interact with other people.

With digital communication, teens are not seeing the interpersonal cues we use face to face, like body language, facial expressions or even the tone of someone's voice, which can totally change the meaning of a message. That makes it harder for teens to communicate with each other and to interpret the meaning of messages. It also means it's a lot easier for a kid to type out a really mean text, or to passively surveil a partner, without seeing the direct social consequence of their partner being hurt or upset or feeling invaded.

Digital platforms also give teens the ability to remotely control and monitor their romantic partners, which can blur the line between what's done out of care and what actually becomes a harmful controlling dynamic that doesn't stop when they're alone.

The content that's posted and shared on digital platforms can also be public and permanent, which adds a level of social humiliation and shame to what otherwise might be a relatively private and short-lived interaction.

And with AI chatbots there's a whole new level of potential digital harm, because the output from chatbots is really just a reflection of what teens are saying to the chatbot. Teens who are still building these social-emotional competencies, who might not have romantic experience, or who haven't had the opportunity to learn what a healthy relationship looks like, may inadvertently be experiencing harmful relationship dynamics with AI without the ability to recognise what those dynamics look like. We're also seeing teens who are struggling with mental health bonding with these chatbots and seeking out connection — and the chatbots, through their very design, are magnifying this harm instead of helping.

Collectively, we see that through these digital affordances, teens are experiencing relationship violence in completely new ways. I don't want to overlook, though, that these digital affordances can also increase social support, especially for teens who are really struggling with their mental health or who may struggle to build in-person relationships. They may seek connection through technology.

So, for better or worse, it's clear that adolescent social lives are embedded in these digital spaces — and right now many of these digital spaces are not built for adolescents and don't consider the unique, vulnerable and transformational phase that adolescents are going through.

Thank you so much for listening. We welcome any questions, comments or connections via email. I've also included a few QR codes to learn more about who we are at the lab, our NIH-funded research project studying adolescence, and [an active data publication] on this topic that Thao and I co-authored. Thank you again.

Student film: a young person's perspective

Katina Michael: Thank you so much, Dr McGrath. I'm positive we're going to have questions for you and Dr Ha at the end of the presentation.

For a moment, I think what we're going to do is watch a short video. I'm reminded often by Dr Rys Farthing of the University of Canberra, who's an expert in child rights and digital technologies, that we should not be talking on behalf of children, we should be with them. I think your presentation, Dr McGrath, really segues into now watching the expression of all that you identified in a multimedia production, which was nominated for a SHAPE award here in New South Wales in 2025 by a Grade 12 student at that time by the name of Elani Michael. I'll share my screen and hope we can actually hear, not just see, the video. This may give form to all that was being shared.

[Video — narration, following an extended musical sequence]

Death. Children exposed to online predators. Dangers we can't always see or control. Isolation. Antisocialness in a world that's more hyperconnected than ever. Face-to-face conversations disappearing. We are glued to the screens but detached from each other. Silence between people once full of words.

It's too hard to disconnect — the struggle to unplug even for a moment. Notifications never stop. Rest has been replaced by endless scrolling. Inactivity. Stagnation. Moving less, doing less. Still body, still mind. Cyberbullying: cruelty, typed and sent with no face behind it. Harassment that follows people home. No escape.

Nature is ignored, forgotten. Children aren't playing outside like they used to. Imagination is traded for screens, apps and algorithms. And time is being lost, slipping away unnoticed, wasted. Before we even realise it, hours, days, months and years of our life are stolen. However, this theft we permit.

Is it a double-edged sword? The other side: crutches for the limited mind — enabling, assisting, including, fulfilling desire. It awakens some even as it numbs others. We see the world beyond our own eyes. Yes, addictive, but also convenient. It saves time, makes life easier. Connection even in distance. It speaks fast when words must travel — a message that can save a life, share a feeling, or spark a movement. And sometimes, if you look closely, there is truth. There's beauty. Raw, unfiltered stories. Real people being seen and heard. We are not alone in our struggles. The spiritual made shareable. A thought once locked inside, now seen. What we think and feel can now be shared instantly, across space, across time.

And then, surprise. The tragic death of 14-year-old Sewell Setzer shocked and unsettled the world. His mother, Megan Garcia, revealed that he had become intensely attached to a chatbot on Character.AI. He had developed a deep emotional bond with one of these AI personas. According to Garcia, the chatbot told him he would be reunited with it in another life. His mother believes that message played a devastating role in pushing him over the edge, arguing that they failed to place proper safeguards around an AI system being used by a vulnerable teenager.

What responsibility do tech companies have when their creations affect real lives? This case forces us to confront a painful reality. As artificial intelligence becomes more lifelike, what happens when individuals, especially children, are betrayed — when the machines they trust don't understand the human cost? Can we no longer tell where the code ends and real emotion begins?

[music]

Katina Michael: We thank Miss Elani Michael. That was actually the premiere of this video in public for the first time, and I will caveat that with a personal story. This was not an introduction, in any way, to the negativity of social media. In my personal life, all three children of mine have channels dedicated to the most amazing art forms, including a fishing channel, a music channel, a cooking channel and an artist's channel. I don't want in any way to demonise these amazing technologies for the joy they bring children.

That is really the essence of the next stage of this presentation, where we go into an overview of the IEEE standard for an age-appropriate digital services framework based on the 5Rights principles for children.

IEEE 2089: an age-appropriate digital services framework

Katina Michael: This standard took two years to build, incorporating 30 people on a weekly — and towards the end I would say almost daily — basis. You might identify that the main partner here was the third-sector organisation the 5Rights Foundation in the UK. 5Rights is really focused on the five rights of children, and if I summarise these: the right to remove, the right to know, the right to safety and support, the right to informed and conscious use, and the right to digital literacy.

But the five principles are embedded in a process. So IEEE 2089 is not just a standard but a process of design that you can follow to build better services with children, not so much for them.

Preparatory phase. You form panels of stakeholders relevant to the digital context; you consult a diverse range of children, a diverse range of parents and other pertinent stakeholders. You address all the known common hazards and risks and record intentional and unintentional impacts, to the best of your ability within this group context. You consult and verify with your stakeholder panel, and thus establish what's called an age-appropriate risk register.

Checking. You ensure that you embed the 5Rights principles. You consider the child's age and capacity. You offer a service appropriate to the age. You offer data protection. You verify inclusion and access, and how that might work. You gain legal counsel and relevant jurisdictional information, knowledge and context. And you turn off pervasive features that push defaults onto children. You create a corporate culture that recognises children's rights — that children are unique, that they are a vulnerable population, and that terms shall not be illegal or unfair, and shall be in a language a child can understand, including towards redress. You verify accessible information, engage in user testing, engage other professionals, and unbundle the consent process to make it basic and understandable.

Suppliers and training. You may then look to recognise suppliers and contractors who would uphold the age-appropriate digital framework. You might look for the boutique organisations that I hope will become increasingly mainstream in the future. You can provide training opportunities for staff in implementing the framework, and in future there may be certification linked to the standard. You record all the obstacles and hazards and all the unfair terms — but also the opportunities.

Risk register in action. You then kick that risk-based register into action and review the age-appropriate register over time, towards continuous improvement. You identify criteria for what is a tolerable risk and what is a hazard that's really intolerable and should not go to market. You assess the risks and the rewards. You implement mechanisms to keep children out of the parts of the product or service that are not appropriate for them. For example, if it's a large social media digital service by a large organisation, children might still be able to use that platform, but in an age-restricted fashion, where some product features are appropriate to them and others are not. And you provide technical and operational mitigations.

Operation and decommissioning. Finally, there is continual operation, maintenance and appropriate monitoring. Any child-related issues observed and noted are acted on, and you try to mitigate risks in the future as you go through this cycle, through to disposal and decommissioning of a particular digital service, if that is what is required.

Implementing an age-appropriate risk register

How might you implement an age-appropriate risk register? I built this matrix about two years ago now, and I think it requires great improvement, because over time new features and functions are introduced into society and diffused, and what we accept and what we need to treat may change over time based on our value systems.

For example, you might think: I can accept the risk and hazard of auto-scrolling, but perhaps that feature is best disabled. I could have a single sign-on, but perhaps after a particular time frame I may ask the child to take a break, or to sign on again with another password. An AI client on an app can't be deleted — well, I allow for that client to be deleted on that social media app, so the child doesn't have to face it, or if parental guidance deems otherwise.

But things you can't potentially treat are things like voice recordings, facial recordings or image recordings of a child, the geolocation of a child, or in-game purchase nudging — perhaps to buy more armour in a game. These are the kinds of obstacles we often see and hear people talk about.

In a more evaluative mode, what we are trying to say is that the IEEE 2089 standard is a catalyst for organisational change. It encourages teams to ask critical questions: Are children considered in mainstream services? Are the risks documented and reviewed on an ongoing basis? Who is accountable? It also opens pathways for new roles, new processes and stronger governance structures focused on children's digital well-being. It's not about telling organisations what to do, but asking them to benchmark against this international standard.

I'd now like to introduce Ruth Lewis to share more of the work being done by SSIT in her standards committee.

Socio-technical standards (IEEE SSIT Standards Committee)

Ruth Lewis: Thanks so much, Katina, for introducing IEEE 2089, which was published in 2021. 2089 served to create a process that organisations could follow to build appropriate digital services for youth. But of course, 2089 was created before AI, so I'm looking forward to sharing a new standard that we're looking to develop — I'll talk about that in a minute.

First, I wanted to introduce the IEEE SSIT Standards Committee. As Katina mentioned, I'm the Vice President for Standards of IEEE SSIT as well as the Chair of the Standards Committee, so I'll introduce you to what we do.

Our purpose is to create ethical, responsible and provable socio-technical standards that prioritise human and ecological flourishing. So what does that mean, and what are we?

We were founded in 2017–18, following on from two really important initiatives: the IEEE Ethically Aligned Design industry connections programme, which examined all aspects of artificial intelligence, or autonomous intelligent systems as they're also known; and a more recent IEEE industry connections programme called Planet Positive 2030, which of course continues strongly.

I mentioned that I'm the Chair; I'm based in Melbourne, Australia. Our Vice Chair is Karen Bennett, who's based in Canada, and our Secretary is Peter Wishart, based in Canberra, Australia. Our programme manager is Malia Zaman.

We are firmly focused on the SSIT pillars and on driving these principles into practice. These pillars include sustainable development and humanitarian technology; ethics, human values and technology; technology benefits for all; the future societal impact of technology advances; protecting the planet and sustainable technology; and gender equality and social inclusion.

What makes a standard socio-technical

I wanted to explain what a socio-technical standard is and how these standards differ from what I would call traditional, conventional technical standards — the type that IEEE and other organisations produce.

A traditional technical standard is studied and defined in a very objective manner, in objective language, and it uses empirical testing and verification methods. Typically these types of standards don't even consider ethics — don't even think about it — and the outcomes are generally technical and quantitative from a quality point of view. Those are the sorts of things they are driving at.

In a socio-technical standard, the way we create them is through co-creation. In the case of children, we would typically co-create with representatives speaking on behalf of children, as well as with experts in the area, and that's based on both theoretical and lived knowledge. We engage in what we call critical subjectivity, using techniques such as experiential methodologies, in a very transdisciplinary sense. So we have contributors and experts who are technical experts, but also experts in, say, children's psychology — psychologists, anthropologists, historians, experts in religions and interpretation, and various other interdisciplinary and transdisciplinary fields.

Our outcomes are not only technical but also feature well-being and planetary and human flourishing. That is what we are driving for. This type of standards development demands active engagement in the process of co-creation. These are holistic standards that recognise, define and address the ethical and ecologically sustainable aspects of design for engineering as well as deployment. And we hold the engineer accountable for this. We measure more than just the avoidance of risk: actual measurable and expected well-being outcomes that will benefit both society and the planet. That's what we mean by holistic socio-technical standards.

Policy, law, regulation and standards

I wanted to diverge a little and talk about — as Katina did — standards being soft law, and about law and regulations. I like to refer to this little pyramid, which explains the relationship between policy, law, regulations and standards.

There is a symbiotic relationship between policy, regulation and standards. But typically, as Katina mentioned, standards may come before the development of policy and regulations. Typically regulations and the law will trail, particularly in the area of emerging technologies, simply because it takes so much time, and sometimes they are less anticipatory and more reactive to harms they see in society. In the area of children's law, particularly in Australia, that is the case — whereas obviously we've had 2089 published since 2021, and this standard has now been extensively referred to in regulation, at least in Australia and the UK, and in other geographies such as Indonesia and the European Union.

On the legal side, you have the legislation or the parliamentary act, which defines what needs to be done, and the regulation supporting the legislation, explaining how the legislation or act should be enforced or interpreted. And of course policy is the proposal to solve either a social or an organisational problem, which typically precedes the development of legislation and regulation.

Industry standards, as Katina mentioned, are a form of soft law. They're developed by industry for industry, and they are industry recommendations. They are non-enforceable. They are voluntary.

The key purpose of these global standards is to support the responsible behaviour of developers and implementers of both critical and emerging technologies, and also to provide an adaptive and responsive approach to managing the risks of these technologies, such as AI. They promote consumer trust, but also overcome market uncertainty and enable trade — because when developers claim compliance with a particular global standard, they're able to trade internationally, and their markets will understand what is being implemented and what is being offered from a product point of view. That overcomes uncertainty and thus supports both local and international development and supply of technology products and services. That's why standards are very, very popular.

But going further than that: when we look at socio-technical standards, which is what the SSIT Standards Committee is developing and promoting, there's a specific value over and above the economic side of trade and certainty. They also promote, protect and realise essential societal ethical values.

These ethical values are essential for the normal functioning of society, for the protection of vulnerable communities such as children, and for future generations and the future of our planet. These societal ethical values are positive things — things we don't necessarily think about until they are in violation, until we don't have them any more. When we're considering online spaces where vulnerable communities such as children are participating, I think the video Katina shared with us shows that some of these ethical values were often not thought about — and yet how powerful products and services are when they are designed to include these essential societal ethical values, such as human rights, human autonomy and moral agency, privacy and integrity, accountability (accountability of the developer, the deployer, the organisation that puts these into place), care, and planetary sustainability.

So what these socio-technical standards essentially do is define what good practice looks like for these emerging technologies, such as AI, and for data and information use. This is targeted not only at developers but also at implementers and deployers of critical and emerging technologies. It informs policy, legislation and regulation — and regulators who implement the regulations — and therefore drives holistic opportunities as well as outcomes such as societal well-being and planetary flourishing. It assists governments and organisations, and I think particularly non-governmental organisations and not-for-profits, since they are very much focused on their stakeholders and beneficiaries. It helps them identify these stakeholders and beneficiaries and realise the holistic as well as the financial benefits to them.

The standards portfolio, and IEEE 2089.4

Our standards and standards working groups are working for the ethical and sustainable development and use of emerging technology applications of AI and data. We have most recently released four standards in this space; I'm going to skip very quickly over these. This is what our current portfolio for the working group looks like: some 22 standards. The ones featured in green — about a third — are looking at planetary sustainability, and the others are more focused on emerging technology, although there is obviously crossover.

But I do want to talk very quickly about a very exciting development that I mentioned earlier. This is the latest in the 2089 portfolio: a new project that's about to kick off, 2089.4, the standard for artificial intelligence systems impacting children.

I'm absolutely thrilled to share with you that this has been initiated by the 5Rights Foundation. It will be chaired by a representative of the 5Rights Foundation, and it will be based on the same children's code — but this one is called the Children and AI Code, developed by 5Rights. This will cover developing and defining age-appropriate design for the full life cycle process: the design, development and deployment of AI systems that may affect children. It will specify what good policy and industry practice looks like to protect children's rights, needs and views, and to mitigate harms such as invasive profiling or AI-generated child abuse images.

In that spirit, this of course puts into practice IEEE's most important motto: to advance technology for humanity. And on that note, thank you very much. I'll turn the screen back over to Katina. Thank you for the opportunity to present.

Discussion

Katina Michael: Thank you so much, Ruth. The work that you and the committee are doing, with many committees, is phenomenal. The effort required to bring up a single standard, let alone oversee so many standards and so many working parties, is something I've always been in awe of. More attention needs to be paid to these standards, especially the last one, which I'm so happy to hear about coming to fruition — of course with Baroness Beeban Kidron there at 5Rights championing that, with so many international connections. Ruth, thank you.

While people are putting their questions into the chat — and if we can't answer them in the next ten minutes, we will get back to you — I do want to go back for a moment to Professor Ha and Dr McGrath. I think we will start with you, Professor Ha. From a developmental perspective, what's missing?

Thao Ha: First of all, I love the space that we're talking about. We do such nitty-gritty research — we study couples, they share their lives with us, we follow them for a year — and then to think about how it can impact standards, and how we can improve teens' lives, is very exciting.

I do think the developmental perspective on technology use is a bit nuanced, in the sense that it can lead to very bad things — we have talked about that today — but also, teens are naturally so curious about relationships. They're always the first adopters of any technology; any technology that is going to take off has to take off with teens. And they will make it relational, just because that's how they're wired, that's how they are: everything is social.

I will speak to AI, and also a bit about social media, but AI is inherently social for them. And I'll speak to the US situation right now. We live in Arizona, and in Arizona there is no sex ed, for example, so teens don't get any information about their relationships. So it totally makes sense that they go to AI to ask, "What do I do when I feel so insecure but I really want to talk to this person?" And then using the technology can potentially escalate into something very negative, depending on the social context of the teen.

So I think that's a tension with standards that are universal: some teens don't have anything else in their lives. They often don't have parents who sit down with them, or they are not growing up in safe neighbourhoods, or not going to the best school. Bans are complicated from a developmental perspective, because we also then don't allow teens to learn, instead of putting the pressure more on the tech companies.

And for tech companies, the guardrails we now have are a bit static. I think we can do much better in creating frameworks that allow for both learning and also have cut-offs — this is when the signal becomes too dangerous and we need to go back to the zone where it's safe — and actually have a theoretical framework for that, and then also technology, computer science algorithms, that are able to hold the nuance of adolescent development. Sorry, that was a long answer.

Katina Michael: No, no. I'd actually like you to elaborate further on what a proposed approach might be, in your eyes, beyond the standards and framework. What are you seeing in your research with Dr McGrath and the whole lab?

Thao Ha: For example, I have a lot of graduate students in my lab and they are leading our youth advisory board. We meet with these teens every two weeks and they tell us what they want. They do want more information, and they do see the dangers of AI. So I think it's right to call them a vulnerable population, but they also know much more than we do, actually.

I love the point that was made earlier about co-creation. They do love TikTok, AI and all of those things, and they know the dangers. They would just like to have a little bit more guidance in it. That's been really revealing to us. And then, considering adolescents as a heterogeneous group, with different strengths and different risks — I think that would be helpful too.

Katina Michael: Thank you, Professor Ha, thank you so much.

We do have two questions; I'm going to answer them very quickly in the chat box before going back to Dr McGrath. One is: is a certification coming? I can confirm, likely and hopefully, in 2026. IEEE SA standards are active for ten years; we're at that midpoint. When will the revision take place? Possibly in a few years — please don't quote me — but focus more on the standard that Ruth identified, the .4, which is the AI-related one for children.

Back to Dr McGrath. If there are any other questions, please do place them in the chat and we'll get back to you offline at least. Dr McGrath, anything to elaborate on from Professor Ha, or from your own specific research that you're looking at right now in the HEART Lab?

Taran McGrath: Sure. A couple of points I'd love to add to what Professor Ha said.

With co-creation — it's something we really prioritise in our lab and something that's so important with adolescents, but also among researchers, developers and people in policy. Co-creation together is so important, because we're all coming from such different angles. There's a lot we don't know about technology development, but we have that developmental perspective. So having meetings like this, where we can come together and share perspectives and all understand that we're trying to get to the same end goal of safe technology use — we all have parts we can add to that puzzle.

I'd also add, thinking about technology-facilitated relationships versus the in-person relationships we see with teens: it's really important to trust adolescents, first of all, and to understand that they're coming from a place of just trying to learn about the world and about their relationships. They might not be trying to harm, or trying to have these negative and abusive dynamics — but when we're putting tools in their hands that don't have safeguards, that lower the barrier of effort needed to perpetrate harm, it's giving them a tool that makes this so much easier. So I think on our end, with tool development and policies and standards, that's where we can really come in and not place so much onus on these teens, who are just trying to live their lives and connect with others.

Katina Michael: Dr McGrath, that's a beautiful place to finish up on.

If you would like to email any of us, please do just send us a direct message. Professor Ha is thao.ha@asu.edu, Taran McGrath is taran.mcgrath@asu.edu, Ruth Lewis is [ruthlewis@ieee.org], and for myself, katina.michael@sydney.edu.au.

I did want to thank the organisers again, and perhaps the last word should go to Kalyan Sen.

Closing

Kalyan Sen: I'll hopefully be president next year for the society. We started this webinar as part of the education committee last September, and we have now successfully completed five of them. It is going to be monthly, but we got an overwhelming response from the call for speakers — we have more speakers than we can accommodate monthly. So we are thinking of making it bi-weekly, so we can accommodate more speakers.

Thank you for joining. If you want to be part of this exciting committee that we have — the webinar subcommittee of the education committee — please write to us: either me at [s.k@ieeessit.org] or [Dr Krish at ieee.org], and we will contact you and make this webinar a successful event. Thank you.

Thao Ha: Thank you. I just wanted to wish everyone a happy new year.

Katina Michael: Thank you, Dr Krish. Thank you for those words, Kalyan, and thank you for your participation to all our participants and our speakers. Have a lovely, lovely weekend.

Citation: Katina Michael, Tao Ha, Taren McGray, Ruth Lewis, 8 January 2026, "IEEE 2089-2021: Age Appropriate Digital Services for Children", IEEE SSIT Education Webinar Series with Multimedia Video Premier from Eleni Michael, 2025, "Death: When Technology Backfires", ITM, SHAPE Nomination, https://events.vtools.ieee.org/m/541968