More US teens are using AI for studying. Not all of them feel great about it.

With investors’ opinions about the future of AI diverging, new research suggests that views about the tech’s impending outcomes are also mixed among some of its youngest users.

A new survey from Pew Research Center, published Tuesday, asked US teens aged 13 to 17 about their AI use, and found that more than half (54%) reported using chatbots to help with schoolwork in 2025 — second only to searching for information (57%) as the top reason, and ahead of using it for fun or entertainment (47%).

Overall, the share of American teens who report using AI for studying has risen significantly in just a few years: back in 2023, Pew found that 13% of students said they ever used ChatGPT for schoolwork; in 2024, it was 26%. Now, only one year later, that share has more than doubled.

Still, even as more teens are turning to chatbots for help researching, editing, or problem-solving, that doesn’t mean they all feel entirely happy about the overall impact that AI might have in the coming decades.

While almost a third of respondents said they think that AI will have a positive impact on society over the next 20 years — with 30% of that cohort citing that it “makes life better, easier” as their top reason — there was still a 26% share that said it would have a negative impact.

Why? Well, besides a history of AI-linked mental health crises, and ignoring the countless hours spent speaking to imaginary characters, the top reason listed in the survey was “overreliance, loss of critical thinking or creativity.” With student test scores slumping in real time, and 1 in 10 teens reporting using chatbots for “all or most” of their schoolwork, could those concerns already be coming into view?