Sonia Livingstone is an international expert on children’s rights in digital contexts, and one of the 40 members selected by the United Nations for the Independent International Scientific Panel on Artificial Intelligence (AI). It’s the world’s first global scientific body on AI, which will assess how AI is transforming our lives.
With one in three internet users across the world being a child, Livingstone will use her expertise to ensure children’s rights are heard in a world where, as UN Secretary-General António Guterres recently underlined, “AI is moving at the speed of light.”
AI creeps into our online lives
In the last few years, AI-based tools have emerged in our daily digital lives, and are now integrated into search engines, social media platforms, messaging services and apps.
Children are therefore also exposed. This can be through their own smartphones or tablets, or through their parents’ and teachers’ use of AI at home and in schools, often for research, educational purposes and games.
“Children are curious and excited, and want to experiment and explore what AI can do,” says Livingstone, who is Professor of Social Psychology at the London School of Economics and Political Science (LSE), and her research is part of the Digital Futures for Children centre.
Many children are growing up in a household where parents speak to an AI companion, such as Alexa, Siri, and Google Assistant, as if they are a person.
Children are particularly fascinated by Generative AI (GenAI), which uses machine learning to rephrase, adapt, and analyse existing data, generating new data that have similar characteristics. Some use it daily, for example when doing their homework or choosing what to wear.
A recent report by the EU Kids Online network, which Livingstone contributed to, found 72% of children aged 9-17 in Europe are GenAI users, although use varies across countries. There was a near-universal level in Austria (94%), and very high levels in Belgium and Italy (89%), Serbia (88%) Portugal (85%), and Croatia (79%), and markedly lower levels in Ireland (40%), Spain (47%), Switzerland (53%), and Norway (59%).
Too early to assess AI’s full impact
Whilst children can benefit from some of the wider advantages of AI, such as improving the efficiency of services, or helping us better understand our planet, the breakneck speed at which it has entered our world means they are also exposed to its risks.
It is too early to know the full impact of AI on children, such as on their sleep, or if AI tools are more beneficial compared to non-AI resources. AI-assisted learning has been shown to overcome disadvantages posed by various disabilities and special needs, Livingstone notes, although she adds that she wants to see further research on this.
“Is it changing children’s understanding of what human is because they can now talk to companion AI? Is it damaging their mental health in ways that will show up in 10 or 20 years’ time? These things we don’t know,” she warns.
Livingstone has not seen a decline in critical thinking or expression among young people in recent years, “let alone a decline attributable to uses of AI.”
“I do see children becoming more cautious and distrustful about what can go wrong with technology, and I anticipate that they will find some new ways to learn, think and express themselves in a fast-changing digital environment, whether for worse or, hopefully, for better,” she adds.
Serious risks facing children
One of the key threats from AI is technology-facilitated child sexual exploitation and abuse.
“That’s probably one of the most shocking and visible forms of harm, and it takes the form of AI-created sexual abuse content and sharing nudification [using AI and deepfake technology to make a person appear nude] apps, new ways of using AI both to approach and to exploit children,” says Livingstone.
Organisations that deal with technology-facilitated child sexual abuse and exploitation are reporting increases in the circulation and creation of illegal content, she adds. Fears therefore abound that there is an increased number of victims.
A further challenge is how Large Language Models (LLMs), such as ChatGPT, Gemini and Claude, are “hoovering up” children’s data “without consent, without consideration and without attention to the fact that this comes from children,” warns Livingstone.
Children are often not anticipated as users of digital resources, Livingstone underlines, and so their needs and rights are often not supported. She is also concerned that AI tools are not being developed by educationalists or child specialists.
“It’s promising, but I’m not going to say it’s beneficial as of yet.”
International expertise for a global challenge
The 40 experts – 19 women and 21 men – on the AI scientific panel, which is supported by the UN, were chosen out of a global call that drew more than 2,600 applications. Experts were selected for their expertise in AI, and geographical representation and gender balance also came into play.
The panel, the UN Secretary-General has said, will be the “first global, fully independent scientific body dedicated to helping close the AI knowledge gap and assess the real impacts of AI across economies and societies.”
It has already held its first meetings and divided itself into seven working groups. It will prepare an initial report for the AI for Good summit in Geneva in July, organised by the International Telecommunication Union (ITU).
Livingstone’s working group focuses on human rights, information and democracy, and she wants to bring in particular her insights on children’s rights.
“It’s super important for a child rights approach that we consult children and young people globally, […] making sure that their voices are heard in ways that are significant and consequential.”
She wants to achieve “what I think the UN is good at, a truly multidisciplinary and multinational, global conversation that achieves improved mutual understanding.”
From Big Tech, Livingstone has heard phrases such as “move fast and break things,” and “winner takes all.” “All of these phrases are very problematic for a human rights agenda, for an agenda of equity, for the SDGs [Sustainable Development Goals].”
She calls for greater clarity, and recognition of how the challenge of AI looks from different parts of the world, and different segments of the population.
“Things seem to me pretty intensely dystopian and anxious in some circles, and intensely pro-innovation for innovation’s sake.”
Navigating a new world as parents
Many parents do not feel equipped to understand AI, or are unaware when their child is using AI, or just how much.
The UK government released in March its screen use guidance for children under five, and Livingstone was on the scientific review panel. It states that “parents should not let young children use AI tools, toys or chatbots (even those aimed at young children) until the present state of knowledge improves.”
Whilst the guidance is realistic about the fact that there will be screen use, with 98% of two-year-olds watching screens daily in the UK, it seeks to limit their time on them.
“There are many other things that children would benefit from doing in terms of social interaction, physical exercise and play,” says Livingstone.
As AI is developing so fast, there are currently no toolkits for parents, which Livingstone calls a priority.
Her advice for the time-being: “Talk to your children, talk with your children, listen to your children, and make sure you are the person that your child wants to come to when there’s a problem.”