MIT professor says universities must rethink credit and training as AI reshapes research

MIT professor Sasha Rakhlin argues that universities should rethink how they train and credit researchers as AI takes on more academic work. His proposals range from teaching students to audit AI-generated results to building shared research infrastructure that records failures as well as successes.

Watch Desk analysis

What happened

In an essay published by MIT News on 8 October, Rakhlin, director of MIT’s Statistics and Data Science Center, sets out questions for universities navigating AI in research and graduate education. He points to mathematics, where AI can produce research results and formal proofs can be checked automatically, as an example of how quickly capability can advance when verification is reliable.

Rakhlin says a polished paper may increasingly be a weaker signal of an individual researcher’s expertise. He argues that departments should recognise contributions such as asking good questions, replication, synthesis, informative negative results and shared datasets, while making clear what researchers contributed and take responsibility for when AI does substantial work.

Why it matters

The proposals reach beyond academic rules about authorship. Rakhlin warns that handing routine calculations, coding and failed approaches to AI could deprive students of the formative work through which they develop intuition and judgement. He says students should learn to formulate problems, audit model outputs, reproduce results and defend their choices.

For research infrastructure, he proposes connected systems that record hypotheses, interventions, outcomes, failures and interpretations, with appropriate permissions. He also calls for investment in computing, secure data systems and expertise in adapting AI models. The aim is to help researchers share useful knowledge across laboratories, including experience that never makes it into published papers.

Our read

This is a more useful university debate than the usual choice between banning AI and pretending it changes nothing. Rakhlin offers a practical test: preserve the work that teaches expertise, while making AI-assisted contributions visible and accountable. The harder bit is building systems that share knowledge without making consent, credit and control an afterthought.

What to watch

• Whether universities make AI contributions explicit in hiring, promotion and funding.
• How graduate programmes teach verification and preserve opportunities to build research judgement.
• Whether institutions invest in shared infrastructure with clear rules for permissions, consent and credit.

Discussion spark: If AI does more of the calculations and drafting, should universities reward the final result, the researcher’s contribution, or the learning and verification behind it?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.