First classroom trial of AI for teachers: time saved, student motivation dented
Researchers led by Wharton's Alp Sungu — with psychologist Angela Duckworth on the team — ran one of the first randomized trials of AI in real classrooms: 193 teachers and over 2,800 middle and high school students at a private school chain in Turkey, over 10 weeks in spring 2025. Teachers given a ChatGPT-based assistant tuned to the national curriculum mostly used it to draft lecture notes, assignments and exams. Their students rated lessons less enjoyable, less interesting and less important than the control group's. Average scores didn't change overall — but students of weaker teachers did worse on standardized final exams. The draft paper, 'Generative AI Can Harm Teaching', is not yet peer-reviewed.
Why it mattersThis is the first causal evidence against the comfortable assumption that teacher AI tools save time at no cost to anyone. The time saved by teachers can quietly become a cost paid by students — especially where weaker teachers use AI output as-is instead of as a first draft. As Google, Microsoft and Anthropic race to put AI in front of millions of teachers, the lesson is not 'ban it' but 'train for it'.
Read in the app — free, in 9 languages
Related stories
Machine learning read the shape of sick brain cells and picked out nine already-approved drugs that calmed them down2026-08-21For two years schools tried to keep chatbots out. Now they are handing them to pupils and asking them to catch the mistakes.
2026-08-21Two school years, 18 schools, one AI tutor: the gain was about what practising without it produces
2026-08-20The FDA has cleared 1,357 AI medical devices. Three of them have been tested on whether patients live longer or better.
2026-08-20So many machine-written proofs arrived that mathematicians built a filing office — and one of its two checks is done by a language model
2026-08-20