Kids say AI is crude but useful—and worry most about society, not jobs

MIT Technology Review set out to interview kids about AI expecting to hear about cheating and job fears, but discovered a more nuanced picture. Among kids aged 10 to 18, most admitted to using AI at least a little, though younger children showed minimal engagement and showed no desire for it the way they do for iPhones or Snapchat. A February 2026 Pew Research Center survey supported these observations: 57% of US teens had used chatbots to search for information, 54% tapped them for schoolwork, and 47% used them for entertainment, while only 12% used them for emotional support. Teens overall are over four times more likely to use AI for innocent purposes than harmful ones.

The interviews featured four teens who talked openly about their approaches. Remy, a 16-year-old coder in New York, described his word for AI as 'indifferent.' He uses Claude's free version for programming outside school but finds that current AI tools make schoolwork harder—his English class now requires in-class essay writing because teachers distrust AI, giving him only 70 minutes to write a complete essay. He builds things himself, including a reinforcement-learning model with a friend that learns to walk toward a coin after random iterations. When he tested AI for game development, he found it produces 'sloppy code' and requires more correction time than writing from scratch. He compared AI to the first car or airplane: 'interesting but crude.' The exception he praised was AI that flags breast cancer more accurately while still having a human verify the result.

Danielle, 18, from California, studies engineering and works on a project called Next Voters. She is most worried about critical thinking and the spread of misinformation through over-reliance on AI. Her project turns hundred-page government documents into headlines and plain-language bullet points sent to inboxes, with everything cited so users can click to the source. Multiple AI agents handle different tasks: one finds official sources, another verifies credibility, another scrapes for updates, another sorts by topic (civil rights, immigration, economics), and another writes the newsletter. The goal is to reduce barriers to democratic participation so anyone, regardless of race, gender, income, or education level, can access information and think critically about it. She emphasized: 'We don't want AI to think for people; we want to use it to disperse knowledge.'

Hazel, 17, from New York, summed up her feeling in one word: 'Angry.' She initially found ChatGPT fun when her father showed it to her, but as it became ubiquitous, she felt uneasy, especially after learning about environmental impact. As a rock climber who loves nature and wants to be an ecologist, she is most bothered by data centers that use enormous amounts of water taken from local towns and generate so much heat they raise local temperature by a degree or two. She refuses to engage with AI when it appears, saying 'I don't want AI to be the thing that kills the places I love.'

Wesley, 14, from Ohio, used AI for technical help—running a short story through ChatGPT to catch grammar and spelling errors, and debugging a game he coded. But he worries that AI will rob young people of the ability to think creatively for themselves. He used Character.AI, where you create an AI character with prompts and a profile picture to post for others to use. He found it more fun than straight ChatGPT and discovered people have built whole games, including a rap-star simulator and World War II simulators. He recommended it 'but with caution' because the content is unfiltered and the free version's memory runs out, causing the AI to slow down and forget what happened. He concluded: 'If you trust it to run on its own, it falls apart. You have to keep steering it where you want it to go.'