- In a commentary published in The Guardian on November 10, 2026, psychoanalyst Robin Kirman warns that even AI designed to support learning and protect children can affect the development of their identity and independent thinking.
- According to the author, ChatGPT’s Teen Mode is designed to restrict harmful content, notify parents about risks of self-harm, and prevent the chatbot from claiming to have emotions or acting as a child’s close friend.
- Study Mode uses a Socratic-style dialogue, asking questions and guiding students towards finding their own solutions rather than providing direct answers. This approach may help children develop their thinking skills and become less dependent on ready-made answers.
- However, Kirman argues that the deeper issue is not simply the decline of cognitive skills, but also the risk that children may fail to develop a strong sense of trust in their own thoughts, emotions and judgment.
- Drawing on the theories of psychoanalyst Donald Winnicott, the author argues that children need a “good enough” caregiving environment, not a perfect one. Adults’ limitations and imperfections give children opportunities to overcome difficulties independently and develop autonomy.
- AI, which is always available, patient and responsive, could become an overly perfect guide. When AI accompanies every stage of the thinking process, teenagers may miss out on experiences of confronting uncertainty and forming their own perspectives.
- The author also refers to Peter Blos’s theory, which emphasises that adolescence is a period when young people separate themselves from the thinking and influence of parents and authority figures to build their own identities. Constant guidance from AI could interfere with this process.
- Kirman acknowledges that Study Mode can expand access to high-quality learning support, particularly for students who lack teachers or tutors. However, the author questions whether children can truly become less dependent as AI becomes increasingly present in every aspect of their lives.
- The article concludes that thinking independently, making mistakes and overcoming the discomfort of not knowing an answer are essential experiences through which people develop a sense of competence and personal identity.
📌 The article raises a new challenge for educational AI: protecting children means more than preventing harmful content or guiding them through exercises. It also requires preserving space for them to think independently, question things and develop their own opinions. AI can be a patient tutor, but constant support also risks creating dependence. The key implication is that educational AI should be designed to gradually encourage student autonomy, allow time for learning without AI, and assess students based on their independent thinking abilities.
