There is a fascinating contradiction emerging in the way we think about children and technology. We are increasingly comfortable saying that children need protection from social media. At the same time, we are asking schools to prepare them for a world in which artificial intelligence will be embedded in almost everything they do. The latest Ipsos Education Monitor 2026 brings this tension into sharp focus. Seventy-eight per cent of Indians surveyed support banning social media for children under 14, while 57 per cent support banning smartphones in schools. Yet only 26 per cent favour banning AI, including ChatGPT, from schools, while 61 per cent oppose such a ban.As a child psychologist, I don’t find this contradiction particularly surprising. Children do not experience all technology in the same psychological way. A platform designed to keep a child scrolling, checking, comparing, and returning for another hit of social validation is fundamentally different from a tool that can help a student understand a difficult concept, generate questions, or explore an unfamiliar subject. The device may look identical in a child’s hand. The psychological experience can be worlds apart. The screen is no longer the useful unit of measurement. For years, our conversations about children’s technology use have revolved around one question: How much screen time is too much?It is an understandable question. It is also becoming an increasingly inadequate one. Two hours spent watching an endless stream of highly curated videos, checking notifications, and comparing oneself with peers is not equivalent to two hours spent creating something, researching a subject, learning a language, or engaging with an AI tool under a teacher’s guidance. What matters is what the technology is doing to the child’s attention, sleep, relationships, emotional regulation, and sense of self.Social media is particularly complicated because it does not simply occupy a child’s time. It can become part of their social identity. For an adolescent, peer approval is developmentally significant. Social comparison is already heightened during these years. Add an environment where popularity can be measured in likes, followers, and views, and the psychological stakes become considerably higher.Then there is the absence of a natural stopping point. A book ends. A television programme ends. A classroom ends.The social-media feed does not.For a developing brain, which is still acquiring stronger executive functions such as impulse control, planning and delayed gratification, an environment engineered around continuous engagement is a formidable opponent. This is why I support meaningful age-based restrictions on social media for younger children, alongside stronger protections around nighttime access.A child’s bedroom should be allowed to remain a bedroom. Sleep should not have to compete with notifications, group chats, and the fear of missing out. Childhood needs periods of psychological quiet time when a young person is not performing, comparing or responding to an invisible audience.AI brings a very different question into the classroomThis is where the conversation becomes considerably more interesting. We cannot prepare children for the future while pretending the future will not arrive. AI is already changing how information is accessed, how work is produced, and how people solve problems. For today’schildren, learning how to engage with AI responsibly will eventually be as important as learning how to search for information online.If a student uses ChatGPT to ask why they got a mathematics problem wrong, AI can potentially deepen learning.Story continues below this adIf a student asks it to explain a difficult scientific concept in three different ways, it can make learning more accessible.If a teacher gives students an AI-generated answer and asks them to identify its errors, the technology itself becomes an exercise in critical thinking.That is an opportunity we should not dismiss. The danger is cognitive passivity. There is developmental value in struggling with a difficult problem. There is value in writing an imperfect first draft, making an error, revising an argument, and discovering that you caneventually solve something that initially felt beyond you. If a child begins outsourcing every difficult cognitive task to a machine, we need to ask what happens to those capacities. We could eventually have students who are exceptionally good at obtaining answers whilebecoming increasingly uncomfortable with uncertainty and intellectual struggle.It is an argument for AI literacy that includes psychological literacy. Children need to understand when technology is helping them think and when it is quietly doing the thinking for them. We also need to stop blaming social media for every adolescent mental-health problem. The Ipsos survey identifies depression and anxiety among the significant challenges facing young people in India, with 30 per cent of respondents identifying mental-health challenges as a concern. The effects of social media and technology, including AI, were cited by 26 per cent.Story continues below this adThese findings deserve attention, but they also require nuance. A young person’s mental health does not emerge from an app alone. Family relationships, academic pressure, loneliness, peer dynamics, socioeconomic circumstances, sleep, temperament and access to psychological support all contribute. Social media can amplify vulnerability. It can intensify social comparison, expose young people to harmful content, and make social rejection feel impossible to escape.But removing an app does not automatically create connection, confidence or emotional security. We cannot outsource the emotional development of children to legislation. Parents still have to create boundaries. Schools still have to teach emotional and digital literacy. Adults still have to notice when a child’s behaviour changes. And parents need to examine their own behaviour. It is difficult to teach children that constant connectivity is unhealthy when the adults around them are checking their phones throughout dinner, conversations and family time.Perhaps the real boundary we need is around childhood itself.The Ipsos findings offer another striking insight. Seventy-two per cent of Indians surveyed believe young people in the country have a bright future – the highest figure among the 31 countries surveyed. Protecting children from harmful technology should not come from fear of the future. It should come from confidence that children deserve the time and developmental space to grow before being exposed to everything the adult digital world has to offer.Story continues below this adAge-appropriate boundaries are therefore essential. A 10-year-old does not need the same digital freedoms as a 16-year-old. A 16-year-old will eventually need more autonomy, along with the skills to exercise it responsibly. This progression is fundamental to child development.At some point, every child has to move from external regulation to self-regulation. That is why a ban can be useful as a protective boundary, but it cannot be the entire solution. If we want children to navigate a world saturated with technology, we have to teach them how technology works on them. They should understand algorithms and attention. They should understand social comparison and digital footprints. They should understand misinformation, privacy, and persuasion. They should know that an algorithm is not a friend, a notification is not validation, and an AI-generated answer is not automatically the truth. We need stronger protections around technologies that exploit immature attention systems. We need thoughtful regulation around children’s access to social media. We need boundaries around sleep and smartphones. And we need schools that teach children how to use emerging technologies with judgement rather than simply banning them out of anxiety. Most importantly, we need adults who understand that technology policy cannot replace parenting, teaching or relationships.The real educational challenge of the AI era may therefore be remarkably simple to state: can we teach children to remain intellectually curious when a machine is always ready with an answer? If we can, then perhaps we will have achieved something more valuable than either a ban or unrestricted access. We will have raised a generation capable of using powerful technology without being controlled by it.