How do U.S. teens actually use chatbots for schoolwork?

Pew Research Center's survey of U.S. teens ages 13 to 17, conducted Sept. 25-Oct. 9, 2025, finds chatbot use for schoolwork and information-seeking has become ordinary: "more than half of teens say they have used chatbots to search for information (57%) or get help with schoolwork (54%)," and "47% say they've done so for fun or entertainment." One-in-ten "do all or most of their schoolwork with chatbots' help," and another 21% do "some" and 23% "a little." Helpfulness ratings run high — about a quarter call chatbots "extremely or very helpful for completing their schoolwork," another quarter "somewhat helpful," and "just 3% say these tools were of little to no help." Against that normalized, broadly positive picture sits a widely shared suspicion about peers: "a majority of teens (59%) think using AI to cheat is a regular occurrence at their school," including "about a third who say it happens extremely or very often."

The survey's own framing links this to a split in how teens weigh AI's effects. Teens see AI's impact on themselves as more positive than negative (36% vs. 15%), but turn more critical "when that same question is framed around society" (negative impact share rising to 26%). Among teens expecting a negative societal impact, the largest group — about a third — cites "overreliance, loss of critical thinking or creativity," ahead of job loss (a quarter) and misinformation (about one-in-ten). So the same population that reports benefiting personally from chatbot-assisted schoolwork also names overreliance and eroded critical thinking as the chief societal risk, and treats peer cheating via AI as commonplace rather than exceptional.

This sits beside the vault's controlled-experiment evidence on the same tension. Does ChatGPT help students code better but remember less? and Does chatbot interaction trade authenticity for better problem-solving? measure the authenticity and ownership costs that Pew's teens only intuit as a social-norm belief ("a regular occurrence at their school") rather than report about their own work. The casual-conversation and emotional-support minority (16% and 12%) also echoes How does emotional chatbot use develop from practical use?, where socioemotional use likewise grew out of practical use rather than starting there, and raises the same companionship concern named in Does sustained engagement with AI companions harm well-being?, though Pew's teens report this use at much lower rates than full companion-app engagement.

The excerpt does not give the survey's sample size, only that it covers U.S. teens 13-17; findings do not generalize beyond that population or country. "Cheating" is undefined and the 59% figure is teens' belief about what happens at their school generally, not a measured incidence rate or a report of their own behavior — it is perception of a peer norm, which can run ahead of or behind actual practice. Helpfulness and schoolwork-reliance figures are likewise self-reported, not measured against outcomes, so they cannot be read against the recall and ownership losses found in controlled studies above; at most, the two pictures run in parallel without confirming each other.