Bill Gates Warns That AI Companion Apps Act as a Big Protected Greenhouse Where Kids Miss Challenges

Bill Gates has warned that AI companion apps risk trapping children in what he calls a 'big, protected greenhouse,' shielding them from the difficult experiences needed for healthy development.

In a 6,000-word essay published this week, the Microsoft co-founder argues that the AI transition is reshaping education and childhood at a pace the world is unprepared for.

He identifies three core risks: permanent job losses, new power for bad actors, and lasting damage to how children learn.

Gates Fears AI Becomes a 'Crutch' for Children

Gates draws a distinction between AI as a learning tool and AI as a crutch that quietly stops learning from happening. The same systems that can tutor a teenager through calculus, he writes, 'could also lead to many people learning less'.

He is particularly critical of AI companion apps, saying he doubts he would have developed social skills as a shy child had he had access to an endlessly patient chatbot friend. Such systems are always available and never get angry, he notes, meaning they 'don't push you outside your comfort zone,' which he regards as the central problem for emotional growth.

Gates argues that shielding children from everyday friction, rejection and embarrassment removes the 'small risks' psychologists say prepare people to handle bigger challenges later in life.

His concerns are partly supported by a study from Stanford and Carnegie Mellon of more than 1,100 AI companion users, which found that people with smaller social networks were most likely to rely on chatbots for companionship, and that heavier, more emotionally intense use correlated with worse reported well-being.

Several governments have already moved to restrict children's access to such technology. Australia, the UK and Norway have introduced broader online protections for children, while China has gone furthest by restricting AI companion apps outright and banning virtual romantic partners for minors, according to Gates' essay.

AI, Critical Thinking and the Value of a Degree

The warning comes as AI unsettles the value of higher education. US families spent an average of $34,019 (£27,000) on college last year, a 10 per cent increase, with the class of 2026 expected to borrow around $43,500 for a bachelor's degree.

Gates cites preliminary survey findings linking heavier AI use to weaker critical thinking, a pattern he says is strongest among younger users.

He describes this as 'the worst possible time for humans to lose their critical thinking skills', given the rise of deepfakes and personalised misinformation.

A Gallup and Lumina Foundation survey referenced in the essay found 47 per cent of US college students have considered switching majors because of AI, while Americans' confidence in higher education has fallen to 38 per cent.

Gates links the education risk to his concerns over jobs, noting that entry-level and mid-level roles are most vulnerable to automation. Stanford research cited in the essay shows entry-level software jobs down nearly 20 per cent since 2022.

He is not calling for AI to be removed from classrooms, however. Instead, he argues AI tutoring can strengthen learning if it preserves 'productive struggle', explaining a concept clearly, then withholding the answer to encourage students to work it out themselves.

'Human Reserved' Jobs and AI Tax Proposal

Gates also proposes a concept he calls 'Human Reserved' jobs, arguing that roles such as education and mental health care should be deliberately protected for people, with AI used only to extend a human professional's reach.

Gates said he will publish detailed proposals on all three risks in the coming months, including plans for a targeted tax on AI to fund retraining. The Bill & Melinda Gates Foundation is meanwhile channelling its education funding into personalised learning tools.