Why the future of connected IoT in homes is child-centric

In 2048 two connected smart refrigerators discuss which one of them is more trustworthy in the homes they serve. The first fridge says: "Children at home trust me. As long as I have something like Nutella in stock, they do."

The second fridge answers: "I am trusted, if the parents at home know I offer the family healthily balanced food, do not share their consumption data without their knowledge, notify them about cyberattacks, have built-in transparency notifications about exploitative consumption-nudging, my AI-enabled appliances run on a stakeholder economic business model, I consume low energy, my real-time data streaming is not impacting children’s development detrimentally and I operate based on a child-centric connected IoT governance framework."

The first smart fridge adds: "If both of our settings are acceptable measures of trustworthiness none of us could be truly trusted. But what are we then, what should we be? Who should we serve?"

Imagining IoT devices to pose self-reflexive questions would require solving decades-old challenges of Artificial General Intelligence which scholars find undesirable. Therefore, the above fictional scenario is only intended to discuss the long-term developmental impacts of connected IoT devices on the youngest generation. The State of the Connected IoT report projects by 2025 41.6 billion IoT devices will capture persons' data with relative continuity.

The rise of an unprecedented industry-unifying connectivity standard called Matter will enable interoperability between different brands and types of IoT devices in homes. Starting a conversation about how the proliferation of AI-enabled IoT devices should shape the youngest generation has never been timelier. Generation Z has already started demanding anonymity now.

According to the State of the Connected IoT report inherent risks of IoT devices pertain to the following: pervasiveness, proximity, granularity, real-time data and fragility to cybersecurity attacks.

As for proximity, the report specifies that IoT devices can be missed as intimate companions. Given that children develop seamless attachment to people, things and devices while growing up in their early lives, connected IoTs in such intimate spheres as their home will only increase the likelihood of their attachments to them.

Looking at real-time data, Anders Raahauge highlights that people "lose the chance to perform privacy checks". Regarding children this poses severe data protection, children's rights, and responsible AI design challenges. Despite the EU high-level expert group's Guidelines on Trustworthy AI advocating for human and children's rights impact assessments before designing AI systems, the proposed European AI Act fails to require developers to implement them. However, without a holistic consideration of the impact of real-time data streaming among connected IoT devices on children’s development, the aggregation of effective practices remains important.

When it comes to fragility to cyber-attacks, – given that IoTs are most prone to cyber-attacks – these devices not only exacerbate the risks to children’s rights online, but also can detrimentally influence children’s moral, mental and civic identity development.

In light of these risks drafting guidance for connected IoTs in homes from a child-centric perspective remains essential. This should entail: a) mitigating IoT risks from children's perspectives, b) mitigating children's online safety risks and importantly c) defining criteria for how child-centric AI-enabled IoT devices should be interoperable for future societies. Age assurance designed in AI systems has been viewed as an effective requirement to protect children from risks. However, it seems to remain insufficient in the age of connected IoTs in homes.

One reason for this is that age assurance requirements in the form of legal codes are powerful but scattered internationally. The European Digital Services Act, for instance, bans adverts targeting children. This requirement creates pressure upon IoT companies to determine users' age, but it has only indirect international reach.

Another reason is that companies interpret age assurance requirements differently. The IEEE Standard for an Age-Appropriate Digital Services Framework is a robust international means to comply with. However, its requirements for IoT companies have multiple bottlenecks. First, no standard guidance exists on how to determine and verify which user is a child and which one isn't. Second, large platforms regularly define the age of users on their own terms. Some introduce sensing AI that harvests behavioral data to determine users' age. Third, the amount of non-intended or erroneous determinations AI technologies could produce remains difficult to foresee.

In the age of Matter, when one smart fridge could interact with another, IoT systems will by default share and analyse pieces of children's daily lives in bulk. The future of designing connected IoT devices for homes responsibly should therefore be with children’s development in mind. This requires a holistic approach toward ideating responsibly about the interconnected impacts of AI-enabled IoTs in homes.

Connected IoT-generated impacts are circular for children. Their usage will shape: a) children’s moral and psychological development; b) how they can generate broadly shared economic prosperity and c) the sustainability of our planet. Dialogue with the meaningful inclusion of youth to define what a child-centric and future-oriented AI governance for connected IoTs should look like is needed.

Methods on pretraining algorithms on synthetic data or measures for real-time data-streaming that involve open source databases such as BLOOM could help better mitigate biases.

IoT devices using AI-platforms that run on a stakeholder economic business model, as provided by Ecosia, could impact children and local communities positively. IoTs could also be built with optical computing hardware to consume low energy. Their cybersecurity protection could benefit from concerted action that is enriched by raising children’s awareness.

Combining similar measures as pillars of a child-centric, future-oriented IoT governance framework can also benefit from UNICEF’s Policy Guidance on AI and Children and the World Economic Forum’s AI for Children Toolkit further. Such framework can yield to youth-informed IoTs and more meaningful youth participation.