MIT report says AI can now complete most university assignments

Artificial intelligence can now credibly complete most assignments set for undergraduate students, according to a new report from the Massachusetts Institute of Technology (MIT).

The finding is based on research at the US university, rather than a study of higher education in the UK. But it raises questions that are increasingly relevant to British universities, where academics are also grappling with how generative AI is changing the way students study and how their work should be assessed.

The MIT report recommends that the university review most of its courses and consider alternative forms of assessment, including oral examinations and portfolios. It also calls for more opportunities for students and faculty to meet in person. MIT president Sally Kornbluth described the report as a “watershed for MIT – and for all of higher education.” For UK universities, the findings underline a challenge that goes beyond concerns about students using AI to write essays or complete coursework.

If AI can credibly tackle a large proportion of the work traditionally set as part of a degree, universities may need to reconsider whether those assignments remain an effective way of measuring what students have actually learned. The report’s recommendations could therefore have relevance beyond the US, particularly as UK universities consider how assessments should work in an environment where increasingly capable AI tools are readily available to students.

The MIT report does not show that the same findings apply to UK universities, but it adds to a wider debate over how higher education should assess students as AI becomes more capable.