Klaus Heine

Mr Heine, what specific challenges arise when children interact with AI?

Klaus Heine: Let me answer this general – but very important – question in abstract terms. Unlike adults, children’s preferences and habits are not yet firmly established. Children are in a phase of cognitive, emotional and social development. This is precisely where their particular vulnerability lies when interacting with AI systems. The risk of addiction and dependency that can arise from AI-supported platforms is therefore far greater for children than for adults. Digital systems react in real time to behaviour, emotions and attention. This not only addresses existing needs but sometimes creates new behavioural patterns and subsequently reinforces them. Children’s so-called vulnerability thus lies, on the one hand, in their developing personality and, on the other, in the architecture of AI-supported digital platforms.

The challenges thus arise from the intersection of human personality development and digital architecture. It is therefore not enough to rely solely on individual media literacy. The technical design of the systems themselves must also be taken into account. This is clearly reflected in recent court rulings on AI and platforms in the United States and in regulatory interventions in the European Union.

What opportunities does AI offer children and young people? And how can these be put to good use?

Klaus Heine: Opinions on this matter vary widely. They range from calls for a largely abstentive approach to AI technologies until the end of school or university, to approaches that advocate the earliest and most comprehensive integration of new technologies into children’s everyday lives. What is important here is that the question of opportunities and risks is not an ideological one, but above all an empirical one. Answering it requires nuanced, long-term and, at times, very detailed research.

At the same time, there is a wide variety of individual opinions. We should therefore be wary of jumping to conclusions. Neither euphoria nor outright rejection will help in developing viable solutions. Rather, the sensible use of AI will be an incremental process of gaining insight. Society, academia, educational institutions and families must gradually learn under which conditions AI can actually have positive effects. In doing so, the focus should not be too narrowly directed at children in developing countries. For there is no doubt that digitalisation and the use of AI offer enormous potential for participation in education and the formation of balanced political views in parts of the world where, for economic or political reasons, access to education is not possible.

How can AI be designed to be child-friendly?

Klaus Heine: This, too, is a very important question, but one to which it is ultimately difficult to give a clear-cut answer. I believe that meaningful answers are most likely to be found in the field of communication theory. Only by combining this with psychological insights and legal regulatory instruments can we develop concepts that do justice to the current digital landscape. I am thinking here of early works by Neil Postman (‘Amusing Ourselves to Death’), who warned as far back as the 1980s that the entertainment industry could become so dominant in our daily attention that we would lose our inner compass for moral action and our ability to assess politics, or that our own behaviour would be subjected to ever-increasing manipulation. This applies all the more to children, who are still finding their place in society, and to AI-driven social media, which is always available.

But even if one chooses not to subscribe to this rather dystopian scenario, media connectivity remains a challenge. This was already succinctly highlighted by Marshall McLuhan (“The Global Village”) in the 1970s. To stick with the metaphor: it is not advisable for children to have access to every house in the global village. Hansel and Gretel would have been better off not entering the tempting witch’s house, where they were to face danger and exploitation. And just as in the fairy tale, parents in the digital world also bear a special responsibility that cannot be delegated to other institutions. Otherwise, what was already touched upon in the first question applies: the creation of addictive potential during children’s formative years should be consistently avoided. The starting point for this is the architecture of platforms, particularly as content providers are difficult to hold to account for a wide variety of reasons. Just think of the right to freedom of expression, the ability to provide material compensation for harm at all, or the enforcement of the law in non-European countries. Conversely, the increasing focus on the digital architecture of AI platforms suggests that, in a sense, it is becoming easier to ensure that AI is designed with children in mind.

The white paper ‘AI and Children: An Overview’ (in German) is available for download here.