A year-long Brookings Institution study concluded that, on the evidence available, generative AI’s potential harms to young people’s learning and development currently outweigh its benefits — flagging risks to wellbeing, to independent thinking, and to equity between schools that can support good use and those that cannot.
The starkest finding is a gap between performance and learning. Students using generative AI were markedly more successful at completing tasks — by roughly 48% in one measure — yet their own unaided performance fell about 17% once the tool was removed. Output rose; understanding did not follow.
Use is already the norm
This is not a hypothetical debate. College Board research found a majority of US high-schoolers already use tools such as ChatGPT for schoolwork. Researchers argue the question has moved on from whether pupils use AI to whether schools can pair it with the teaching — retrieval, explanation, supervised practice — that converts assistance into genuine learning.
Sources
Brookings Institution — A new direction for students in an AI world · NPR — Report: the risks of AI in schools outweigh the benefits · College Board — Majority of high school students use generative AI for schoolwork
