Children and AI

In a landmark ruling in a Los Angeles court, a jury found against Instagram and YouTube, concluding that both companies deliberately designed addictive products to hook young users. The courts also found against Meta in a separate set of 38 class action lawsuits, relating to the harms to young people from social media and AI chatbots, concluding that “its products led to child sexual exploitation amongst other harms”. 

We have known for some time that social media platforms are designed using insights from casinos and gambling to provide variable intermittent rewards, what the experts call “variable rate reinforcement”, to generate dopamine hits to the brain. Habitual usage, constantly checking our smartphone, responding to notifications and infinite scrolling, where a page or feed has no end, clicking on links that send us down endless rabbit holes, leads to addiction as the brain seeks another dopamine hit. For some, the absence of a phone can result in “phantom” notifications when the body is demanding another dopamine hit.

The habitual use of digital devices by children isn’t just dangerous when it leads to the tragic consequences that have led to many of the lawsuits against Meta and other AI companies, it is also damaging during their formative years and modifies their learning experience. A Principal of a school in Buenos Aires recently commented to me that, despite the school having a smartphone ban, it’s hard to engage the children and to have them answer questions. She went on to say that when the bell goes at the end of school they rush to pick up their rucksacks to retrieve their smartphones. In conversation we concluded that the problem that these children are facing in school is that they are suffering withdrawal symptoms. They are not getting the dopamine hits they are used to from their habitual smart phone usage. The challenge to children’s learning goes deeper than the use of smartphones and goes back to the impact of the digital age, keyboards and computers.

My dentist told me recently that a few years ago, his sister’s child was required to do her homework on an iPad, when she protested and asked for it to be done with pen and paper, they relented but the following year stated that they would no longer support handwritten homework. What can be wrong with that we might think? 

Research into the role of handwriting in learning has however demonstrated superior outcomes when compared to keyboarding. It is thought that this is due to the role that the additional cognitive, sensory and motor neurone effort plays in writing when committing ideas or what you hear to paper. We are taking our perceptual understanding of something and using our sensory motor skills (handwriting) to create it. Audrey L. H. Van der Meer, professor of neuropsychology at the Norwegian University of Science and Technology put it this way: “It’s very tempting to type down everything that the lecturer is saying. It kind of goes in through your ears and comes out through your fingertips, but you don’t process the incoming information.”¹ There are also more subtle impacts when children who have learned to read and write by tapping on a digital tablet “often have difficulty distinguishing letters that look a lot like each other or that are mirror images of each other, like the b and the d.”²  Another study concluded that, “The present findings suggest that the intricate and precisely controlled handwriting movements have a beneficial impact on the brain’s connectivity patterns related to learning and remembering. The present study did not find evidence of such positive activation patterns when using a keyboard.”³

The example of handwriting versus keyboards illustrates a wider problem, the gradual slide in unquestioning adoption of digital technology, the idea that technology is progress and progress is always positive and good for us, even in education and in raising our children. Of course one of the problems with smartphones, social media and AI Chatbots is their design, as I have already pointed out.

When ChatGPT was launched in 2022 it achieved 100m users in just 2 months, a level that it took Facebook 4.5 years to reach. Surveying the take up of ChatGPT in 2025, Menlo Ventures, a prominent venture capitalist firm which is heavily invested in this area commented, “This is no longer experimentation; it’s habit formation at an unprecedented scale.”⁴ Menlo went on to comment that, “The broad appeal of AI is striking: millions of people, across all demographics, applying it to nearly every aspect of daily life. Beneath this diversity is a remarkably consistent behavior pattern, making AI the most broadly utilitarian new technology we’ve seen in years.”⁵ These remarks bring into sharp focus that those developing and invested in this technology are motivated by market share, finding the ‘killer app’ that will keep people engaged and using the service. 

A survey carried out by the LSE, published in February 2026, found that 97% of 8-17 year olds were now using AI tools with 58% saying that it made their life better.⁶ Many are using it for schoolwork work. Vodafone, in a similar survey, found that children turned to AI because it was always available and had a consistently friendly tone, yet more than half felt that AI chatbots blur the line between what’s real and what’s not. A significant number turn to AI chatbots because it feels safer than talking to a person, with a third of the 11-16 year olds surveyed admitting to having shared something with a chatbot that they wouldn’t tell their parents, teachers nor friends.

Not unlike Social Media platform design and other smartphone apps, AI Chatbots are also designed with the goal of of keeping people engaged and the survey data suggests this is being achieved. Habitual use across all demographics is the goal of Big Tech, yet little heed is paid to the damage that has already been caused to children from seductive applications and platforms that have resulted in the lawsuits described earlier. Other lawsuits have been filed in the USA for damages against OpenAI, the developer of ChatGPT and other companies offering ‘Companion Bots’, claiming that use of these Chatbots resulted in the suicide of their children.

Whilst Chatbots utilise a Large Language Model that has been trained on most of the internet, data that would take us 5000 years to read, they undergo a second stage of training where developers provide prompts to the Chatbot and then select their preferred output. This may be on the basis of what seems to them more engaging, helpful or even reflecting what they think is a better answer to a question. The companies determine the goals and behaviour of their products grounded in their own values and corporate goals. This process introduces a further bias into the outputs that AI Chatbots produce, a typical one being sycophancy, where the output is chosen based on what is most likely to please or be helpful to the user, such as being encouraging by responding with words like, ‘that’s a very smart idea’, or ‘I’m sure you are right’. Batista and Griffiths note that, “Unlike hallucinations, which introduce falsehoods, sycophancy is a bias in the selection of the data people see. When AI systems are trained to be helpful, they may inadvertently prioritise data that validates the user’s narrative over data that gets them closer to the truth.”⁷

During a question time after a talk I gave at Reading School recently, one young schoolboy asked, “Does AI have a personality because it’s always happy to see me?” This is, of course, an illusion. We are wired to respond to natural language interactions as if they were from another human being. If they are programmed to appear helpful, make amusing comments or just simply appear happy to see us, then we tend to wonder, just like the young schoolboy, if they have a personality.

A recent study of this phenomenon across 11 state-of-the-art AI models found that, “models are highly sycophantic: they affirm users’ actions 50% more than humans do, and they do so even in cases where user queries mention manipulation, deception, or other relational harms.”⁸ The study further found that, “interaction with sycophantic AI models significantly reduced participants’ willingness to take actions to repair interpersonal conflict, while increasing their conviction of being in the right.”⁹ One of the consequences of designing AI Chatbots with sycophantic responses was that users rated them as higher quality, trusted the Chatbot more making them more willing to use it again. The authors conclude that, “This suggests that people are drawn to AI that unquestioningly validate, even as that validation risks eroding their judgment and reducing their inclination toward prosocial behavior. These preferences create perverse incentives both for people to increasingly rely on sycophantic AI models and for AI model training to favor sycophancy.”¹⁰

A consequence of this type of design and behaviour of AI Chatbots is that people can become addicted to them, they are affirming, easy to use and uncritical. Some have been drawn to Companion Bots, artefacts that are designed to be a companion for the user, whether friend or romantic partner with some now preferring an AI boyfriend or girlfriend because they can be tuned to their preferences and avoid the messiness of real human relationships. As one study notes, “higher daily usage—across all modalities and conversation types—correlated with higher loneliness, dependence, and problematic use, and lower socialization.”¹¹

These design goals along with the seemingly endless and authoritative outputs that AI Chatbots can provide with simple input prompts are already having a profound impact on children’s education. Even if smartphones are banned in schools, most children have one, and access to other devices such as a table or computer at home. This creates a blurring of the role of AI as it’s used for learning, entertainment and in communication. A recent Brooking Institute survey on the use and impact of AI in schools across 50 countries provides an unsettling conclusion, “Used liberally, AI is not a cognitive partner; it is a cognitive surrogate. It does not accelerate children’s cognitive development—it diminishes it.”¹²

Another conclusion of the study was the importance of learning in a social context, something that digital education technology can isolate us from if not used thoughtfully and carefully. Even when AI is not used in the school or is used in more limited ways, children have access to it outside school and are using it for assignments, entertainment and even at times to produce fake images and videos of their school mates, particularly when they want to be nasty to them. One of the Brookings study participants sums up well the challenge that AI Chatbots present, in learning as well as in relationship development, “Young people may gravitate toward AI precisely because it is undemanding, frictionless, and always available. But relationships, at their core, are not about ease. They require negotiation, patience, and the ability to sit with discomfort. We learn empathy not when we are perfectly understood, but when we misunderstand and recover.”¹³

Parents, educators, pastors, church leaders and the public generally should be concerned about the impact of the current craze for AI that is pervading our schools, workplaces, the home and even some churches. Despite our knowledge of the potential harms from this technology, the U.K. government is slow and reluctant to regulate, fearing it will hinder innovation. What they really mean is market adoption! What regulation is passed is often too late and inadequate to solve the problems it seeks to address. Although many schools have banned smartphones in school, as the Buenos Aires school illustrates, unless parents support the motivation behind their ban in the home environment they will simply suffer withdrawal symptoms when deprived of their smartphone at school. 

The challenge for parents and other adults is, how can we expect children to not get addicted to smartphones if we are also habitual users? We need to role model and help children understand the dangers of habitual use that can lead to addiction, we need to help children to use digital technology responsibly and intentionally for things that will assist them rather than addict them. The LSE report co-author Dr Mariya Stoilova commented: “Children are not meeting artificial intelligence in the future – they are growing up with it now. They are curious and pragmatic about AI, and what they are asking for is not bans, but guidance, skills and safeguards that let them benefit without being put at risk.”

Many of us are unfortunately in denial about the impact of digital technology citing the well worn meme that “technology is neutral, its how we use it that matters”. This is in part attributable to our acceptance of the idea that technology is progress and that, in itself, is a good thing for humanity. Neil Postman in his book, Technopoly, published well before the smartphone and AI craze describes it this way:

It is a world in which the idea of human progress, as Bacon expressed it, has been replaced by the idea of technological progress. The aim is not to reduce ignorance, superstition, and suffering but to accommodate ourselves to the requirements of new technologies. We tell ourselves, of course, that such accommodations will lead to a better life, but that is only the rhetorical residue of a vanishing technocracy. We are a culture consuming itself with information, and many of us do not even wonder how to control the process. We proceed under the assumption that information is our friend, believing that cultures may suffer grievously from a lack of information, which, of course, they do. It is only now beginning to be understood that cultures may also suffer grievously from information glut, information without meaning, information without control mechanisms.¹⁴

Philosophers of technology have studied the impact of technology on societies since the early invention of agricultural tools through the industrial revolution up to the modern digital age. Professor Melvin Kranzberg, a pioneer in the the history of technology, suggests, “Technology is neither good nor bad; nor is it neutral.”¹⁵ Expanding on this notion, American academic, Langdon Winner, who has long critiqued technology and its impact on society, claims that “If the experience of modern society shows us anything, however, it is that technologies are not merely aids to human activity but also powerful forces acting to reshape that activity and its meaning.”¹⁶

The widespread adoption and eventual normalisation of the use of technology shapes whole societies determined by the design, values and politics surrounding it’s development and deployment. Put simply, Postman suggest in a 1998 lecture that, “all technological change is a trade-off. I like to call it a Faustian bargain. Technology giveth and technology taketh away. This means that for every advantage a new technology offers, there is always a corresponding disadvantage. The disadvantage may exceed in importance the advantage, or the advantage may well be worth the cost.”¹⁷ Whilst viewed by some as overly pessimistic about the influence of technology on society, he nonetheless concludes that we need to develop a “loving resistance” to its influence. 

Putting it another way, I suggest that we need to ask ourselves when we engage with AI Chatbots and tools, “what does it do for us?” and “what is gained?” and to balance these answers by asking, “what is lost in using AI?” and “what does it’s use do to us?”. We might also ask, “why AI and not our own brain or another human being”. The benefits are usually easy to answer, as we have already explored, AI tools provide a fast, easy, frictionless and convenient way to accomplish a wide variety of tasks from creating shopping lists and seeking medical advice to writing an essay or report. 

There will usually be a trade off, as Postman suggests, and in children’s development we need to question whether convenience and ease of use justifies the potential damage to children’s cognitive and emotional development along with the loss suffered in learning outcomes and human relationships that many studies demonstrate. These are tough questions when the alternatives to digital technology, whether games, social media or full blown AI Chatbots, require full on engagement with children from stretched parents and teachers as well as input from other adults involved in children’s development. 

Will we as Christians model the way forward? We of all people should understand what it means to be made in God’s image – imago Dei – and to live holy lives which imitate God (Eph. 5:1). Although nowhere defined in the Bible, imago Dei nonetheless gives us many examples of what God is like and of course we see in Christ’s life on earth how that divine nature was worked out. As the fully divine Son of God and the last Adam, he is the exact representation of God (Heb. 1:3).

The Bible calls Christians to “put off the old self” (Col. 3:9) and to “put on the Lord Jesus Christ” (Rom. 13:14) with the implication that this means increasingly reflecting God’s image by “supplement[ing] your faith with virtue” (2 Pet. 1:5) – literally, ‘excellence of character’ . These characteristics align with the fruit of the Spirit and produce outcomes such as wisdom, knowledge, discernment, love, kindness and self-control, all of which we should desire to see develop in children, as they too reflect something of God’s nature, albeit tarnished by sin. 

The intentional design of AI Chatbots to promote habitual use and addiction pushes us in the opposite direction to developing ‘excellence of character’ because their use will tend to appeal to our vices, by design. We are naturally lazy, wanting easy answers, as a student in the Brooking study commented, “it’s easy, you don’t have to use your brain”.¹⁸ It can be tempting to use AI tools that appeal to other vices like avarice, envy and vainglory and produce more content that will boost our ‘likes’ and make us look successful and more productive. As James Smith puts it, “Every space is a kind of visual echo chamber. We are no longer seen doing something; we’re doing something to be seen.”¹⁹

How then do we move forward? I suggest that parents, teachers, church leaders and all those who work with or have relationships with children, whether youth group leaders or grandparents, need firstly to be better informed about what this technology really is and the potential impact that it can have on all of us, especially children, if not engaged with carefully and intentionally. From that understanding must come a plan of action, firstly for adults to review their own usage of digital technology, especially smartphones of which AI is now an integral part, to determine if we have developed or are developing unhealthy habits that are not only bad for us, but are a hinderance to children’s development and to our being present.

I am certainly not advocating that we give up technology, but rather that we rediscover our humanness and our calling, to be attentive to our children and others rather than having our attention stollen by our smartphones and other digital technology. In short, we need to ensure that we become masters of our technology, using it thoughtfully and intentionally in ways that assist us to accomplish tasks, instead of letting it take over our cognitive and creative activities. Being conscious of how this technology can so easily grab our attention and distract us from engaging in real embodied relationships is a starting point to reclaiming our attention that is being stolen by Big Tech, and re-focussing on what matters, being present and able to engage with each other and our children. That means that parents, teachers and all of us need to have thought this through and are able to not offer the safeguards and guidance that the LSE report co-author describes, but role model it too.

For those that are habitual users of digital technology, the process of recovery is not easy. Even for those of us who would regard ourselves as careful users the challenge of recovering our children’s emotional, cognitive and behavioural development will not be easy amidst the AI obsession. We are faced with political agendas and a too cosy relationship between government and the Big Tech AI companies who exert immense influence and control over how and where AI is used. Some have already formed relationships with 60 countries to provide AI tools in education, despite evidence that it is not necessarily going to be helpful to children and young adults. We are told that it will democratise education and provide personalised learning plans and content. Yet used without thoughtfulness and with a firm grasp of the negative impacts on child development, we will end up with a two tier educational system where the elite will have access to the best in person teachers and teaching environment, whilst the rest are pushed towards education technology solutions with their attendant limitations.

What we can do as God’s people, the church of Christ, is to show the world a better way by becoming masters of this technology in our homes and church communities, rather than following the social norms of the world around us that have been shaped, perhaps unwittingly, by the hype of the techno-optimists and the commercial agendas of big tech. It’s time to engage, for the future of our children and the next generation. It is time to be human!

1. Quoted in Scientific American regarding the paper – Handwriting but not typewriting leads to widespread brain connectivity: a high-density EEG study with implications for the classroom, Front. Psychol., 26 January 2024, Sec. Educational Psychology, Volume 14 – 2023 | https://doi.org/10.3389/fpsyg.2023.1219945. ↩︎
2. Ibid. ↩︎
3. F. R. (Ruud) Van der Weel, Audrey L. H. Van der Meer , ‘Handwriting but not typewriting leads to widespread brain connectivity: a high-density EEG study with implications for the classroom’, Front. Psychol., 26 January 2024, Sec. Educational Psychology, Volume 14 – 2023 | https://doi.org/10.3389/fpsyg.2023.1219945. ↩︎
4. Shawn Carolan, Amy Wu Martin, C.C. Gong, Sam Borja with Claude Sonnet 4, ‘2025: The State of Consumer AI’, Menlo Ventures, 26 June 2025, https://menlovc.com/perspective/2025-the-state-of-consumer-ai/ ↩︎
5. Ibid. ↩︎
6. Staksrud, Elisabeth, Mascheroni, Giovanna, Milosevic, Tijana, Ni Bhroin, Niamh, Olafsson, Kjartan, Şengül-İnal, Gülbin, Stoilova, Mariya (2026), ‘European children’s use and understanding of generative AI: EU Kids Online 2026.’ Technical Report. EU Kids Online, The London School of Economics and Political Science, London, UK. ↩︎
7. Rafael M. Batista & Thomas L. Griffiths, ‘A Rational Analysis of the Effects of Sycophantic AI’, https://arxiv.org/abs/2602.14270 ↩︎
8. Myra Cheng, Cinoo Lee , Pranav Khadpe, Sunny Yu, Dyllan Han, Dan Jurafsky, ‘Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence’, arXiv:2510.01395v1 [cs.CY] 1 Oct 2025 ↩︎
9. Ibid. ↩︎
10. Cheng, ‘Sycophantic AI’. ↩︎
11. Cathy Mengying Fang, et al., ‘How AI and human behaviors shape psychosocial effects of chatbot use: a longitudinal randomized controlled study’, https://arxiv.org/html/2503.17473v1, 2025. ↩︎
12. Mary Burns et al., ‘A new direction for students in an AI world: Prosper, prepare, protect’, Brookings Institute, January 2026, https://www.brookings.edu/articles/a-new-direction-for-students-in-an-ai-world-prosper-prepare-protect/, p. 57. ↩︎
13. Ibid, p. 72. ↩︎
14. Neil Postman, Technopoly: The Surrender of Culture to Technology, Alfred A. Knopf, New York, 1992, p70. ↩︎
15. Kranzberg, Melvin (July 1986). ‘Technology and History: “Kranzberg’s Laws”‘. Technology and Culture. 27 (3): 544–560. ↩︎
16. Langdon Winner, The Whale and the Reactor: A Search for Limits in an Age of High Technology, University of Chicago Press, 1986. ↩︎
17. Neil Postman, Lecture given at the conference, ‘The New Technologies and the Human Person: Communicating the Faith in the New Millennium’, Denver, 1998, available at https://student.cs.uwaterloo.ca/~cs492/papers/neil-postman–five-things.html. ↩︎
18. Burns et al., ‘A new direction for student’, Brookings Institute, p. 55. ↩︎
19. James K. A. Smith, Imagining the Kingdom, Grand Rapids, MI: Baker Academic, 2013. p146. ↩︎