- A Pew Research Center survey found that 54 percent of U.S. teens now use AI chatbots for schoolwork, double the rate from one year earlier when just 13 percent reported using ChatGPT.
- Nearly 60 percent of teenagers say AI-powered cheating happens at least “somewhat often” at their school, with a third saying it occurs “extremely” or “very often.”
- Ten percent of teens rely on chatbots for most or all of their assignments, while 12 percent have used AI for emotional support.
- The survey of 1,458 teens and their parents was conducted online from September 25 to October 9, 2025.
What Happened
A Pew Research Center study published February 24, 2026, found that AI chatbot use among American teenagers has doubled in a single year. More than half of U.S. teens, 54 percent, now use tools like ChatGPT and Microsoft Copilot for homework help. When the same question was asked in 2023, just 13 percent reported using ChatGPT for school-related tasks.
The research team was led by Colleen McClain, Senior Researcher at Pew, and Monica Anderson, Director of Internet and Technology Research. Research Analyst Olivia Sidoti and Research Associate William Bishop co-authored the report. The team surveyed 1,458 U.S. teenagers and their parents online between September 25 and October 9, 2025.
Why It Matters
The study’s most striking finding is not that teens use AI for school. It is that they believe everyone else does too. Fifty-nine percent of respondents said AI-assisted cheating happens at least “somewhat often” at their school. Roughly a third placed the frequency at “extremely” or “very often.” Only 14 percent said students at their school rarely or never cheat using AI chatbots. Another 15 percent said they were unsure whether it happens at all.
That perception matters because it normalizes the behavior. Researchers concluded that teens now view cheating with AI as “a regular feature of student life.” When a majority of students believe their peers are using chatbots to complete assignments, the social pressure to do the same increases, even among students who might otherwise avoid it.
The findings suggest that schools are losing the enforcement battle. Despite widespread concern among educators about AI-generated homework, the rate of student adoption is accelerating faster than institutions can develop detection tools or coherent usage policies.
Technical Details
The survey broke AI usage down by specific task. Nearly half of teen AI users employ chatbots for research, pulling together information from multiple sources through conversational queries. More than 40 percent use AI to help solve math problems. Over a third rely on chatbots to edit and improve their writing.
The 10 percent of teens who use AI for most or all assignments represent the most concerning segment for educators. At that level of dependency, student work may primarily reflect chatbot output rather than individual learning. A quarter of all teen users said chatbots have been “extremely” or “very” helpful for schoolwork, and another 25 percent described them as “somewhat” helpful.
Beyond academics, 12 percent of teens reported using AI chatbots for emotional support, a finding that intersects with ongoing concerns about adolescent mental health and the appropriateness of AI systems serving in a quasi-therapeutic role for minors.
Who’s Affected
Teachers and school administrators face the most immediate pressure. AI detection tools like Turnitin’s AI writing detector have proven unreliable, producing both false positives that penalize students who wrote their own work and false negatives that miss AI-generated text. Parents are navigating a landscape where their children’s homework habits have fundamentally changed in ways that are difficult to monitor from home.
Students themselves are split. Those who use AI tools describe them as genuinely helpful for understanding difficult concepts. Those who do not use them face an uneven playing field when competing for grades against classmates who do.
What’s Next
The Pew data suggests that prohibition-based approaches to AI in schools are not working. Student adoption has doubled in one year despite institutional resistance. The question now is whether schools will pivot toward structured integration, teaching students to use AI as a learning tool with clear boundaries, or continue attempting enforcement strategies that the data shows are failing. The study did not track how school policies have changed in response to rising AI use, leaving a significant gap for future research to fill.
