Proofs and Prompts

Notes from the common room

Recommendations from the Harvard CMSA summit on Math and AI

The organizers of the Harvard CMSA summit1

Last week we organized an intensive two-day summit Math Research in the Age of AI. It was held at the Harvard Center for Mathematical Sciences and Applications (CMSA). Thirty-two mathematicians from academia, government, and industry came together to help navigate this period of rapid change in mathematics and, most importantly, to offer ideas for building the future. There were five working groups organized around the following themes:

  • Evaluation, production, and dissemination of research
  • Coordinated actions for the AI era
  • Open-source tools
  • Formalization in mathematics
  • Collaboration between academia and AI labs

Our report can be found on the summit website. There you’ll also find a link to provide feedback.

Portions of the recommendations, particularly those in the section on evaluation, production, and dissemination of research, offer immediately actionable suggestions for review and hiring in mathematics departments. We encourage departments to act on those they can implement at this time and to begin longer-term conversations to rethink evaluation practices. We recognize that many of these changes may temporarily place greater burdens on faculty. These efforts are necessary to facilitate transitions to new norms for the field and are especially important in light of the need to support our students and young colleagues. Our recommendations go beyond academia; they impact government and industry as well.

Received 30 October 2026.

  1. Dan Freed (Harvard), Larry Guth (MIT), Bryna Kra (Northwestern), Emily Riehl (Johns Hopkins) and Mohammed Abouzaid (Stanford) ↩︎
Filed under
· ·
More by →

New to Proofs and Prompts? Start here →

4 responses to “Recommendations from the Harvard CMSA summit on Math and AI”

  1. Anonymous Avatar
    Anonymous

    That is another pamphlet guiding everyone towards AI.

    But they “tolerate AHM mathematicians” as if they were a weird species and the final section is praising all industry initiatives who bought mathematics.

    That kind of anticipatory obedience can also be seen in some comments here.

    Reply
    1. anon Avatar
      anon

      Indeed. It was difficult not to compare AGMAI with Vichy. But maybe AGMAI was just Bordeaux, and Vichy is the Summit on Math Research in the Age of AI?

      Reply
  2. Zoran Škoda Avatar
    Zoran Škoda

    The recommendations are shallow and also very self entitled. For example when discussing formalization it does not even mention the basic problem that formalization depends on choices, which if done by the tools, often have (and even incentives go into this direction) variant choices and simplification not exactly corresponding to the original version. Or about the problems coming from wrong incentives in the past and aspects of the mathematics which are neither theorem proving nor “understanding” mantra, for example developing new visions and programs (where AI for now did not do much so fat). One prominent mathematician told me so 20 years ago that he spent 90% of time “dreaming” what to prove, how to formulate the dreams and where to go — and only 10% for the actual proofs; now some people talk about making sense of math as a new activity discovered in 2026 as if only the theorem proving existed before. The mantra of dihotomy — proofs versus understanding — seems to be some naive virus from american ivy league math departments whose prominent members so authoritatively promote it. Almost every recommendation in above recommendations has “should” or “must” verb; while anybody who knows mathematicians know that each has so different style mentality, and so on. Not all are collaborative, some are autistic, extreme, erratic, others organized and so on. The awareness of the scales (intermediate term prospects versus long terms prospects of the state of the matter) seems to be absent. The awareness that the traditional mathematics always had so much unpublished culture (folklore, alternative proofs, heuristics, unfinished programs, indviduals in shadow…) and unpublishable culture is missing (while all what is produced from AI is counted, such forms seem to be under the rug).

    Reply
  3. Gordon Royle Avatar
    Gordon Royle

    This is another document full of vague statements long on “should” and “must” and painfully short on “how”. What does “Mathematicians should reshape the profession as it adapts to the rise of new technologies” mean? Do we get to shape our profession at all or will it just evolve organically by the cumulative effect of individual decisions by people each out to maximise their own utility?

    Reply

Comments are moderated. Read our comment policy.

Add to the discussion

Subscribe

New posts by email.

Prefer a feed reader? Subscribe by RSS.

Also on Mathstodon.

Join the discussion

Latest comments across the site.

Read next