MIT released a report in June 2026 that documents how students are using AI tools across the board. The study says fewer students attend office hours, online discussion posts have dropped, and study groups in dorms and libraries are shrinking. Faculty say it’s harder to tell what students actually learn.
What the report recommends
The committee advises against using AI‑text detection software because it often mislabels work by non‑native speakers or neurodivergent students. Instead, every thesis or dissertation should disclose any AI use, and AI may not be listed as a co‑author. Faculty are asked to be transparent when they use AI for slides, feedback or grading.
Instead of blanket rules, the report suggests each course create its own AI policy based on learning goals, assessment design and the nature of the subject. Suggested teaching changes include oral exams, semester‑long portfolios, in‑person discussions and project‑based work.
Wider impact and unequal access
The study warns that students who can pay for premium AI subscriptions – which can cost several hundred dollars a month – may pull ahead of peers who cannot. MIT already gives faculty and graduate students $30 of free AI credits per month through its Parley platform, but undergraduates must buy their own access.
Similar trends appear at other schools. Harvard surveyed students in 2024 and found 87.5 % used AI, with about a quarter saying they visited office hours less often. In the UK, 95 % of full‑time students used AI by the end of 2025, especially those from wealthier households.
Why it matters
Students may skip office hours and study groups, losing chances for direct help and peer learning. Faculty find it harder to assess real understanding, which could affect grades and feedback quality. The gap between students who can afford premium AI tools and those who cannot may widen performance differences on campus.
