An MIT survey published on October 29 reports that a large share of early-career mathematicians have grown less confident about continuing in academic mathematics. The questionnaire collected responses from 268 participants — primarily PhD students and some postdoctoral researchers — representing nearly 60 different institutions.
Respondents were asked how the probability that they would continue in academic mathematics after their PhD/postdoc had changed since mid‑July. Around 70% of those surveyed said they are now less likely to continue on an academic math path, while only 9% said they are more likely to do so.
Key figures from the survey
- 268 respondents from nearly 60 institutions
- ~70%: increased doubts about continuing in academic math
- 9%: feel more likely to continue
- 55% dislike using AI for research, 31% like using AI
- 76% expect it will be harder to obtain a tenure position, 7% expect it will be easier
- 86 respondents would feel “very unhappy” if AI solved a problem they were working on, versus 6 who would feel “very happy”
- 77% feel less motivated to work on math research, 10% feel more motivated
The survey was closed shortly after OpenAI published solutions to 372 families of problems on October 6; the survey authors note that most responses do not take that development into account and that attitudes may have become even more pessimistic since then.
What the mood shift suggests
The results indicate substantial unease among early-career researchers about how rapid advances in AI might affect the value of human mathematical work and the prospects of an academic career. Respondents report feelings of powerlessness, demotivation, and unhappiness: not only do they expect tighter job prospects, but many also fear that AI will displace the research futures they had imagined.
Beyond mathematics, commentators and recent economic papers referenced alongside the survey raise broader concerns. Some economic analyses argue that in an AGI‑driven economy, human wages could become tied to the cost of reproducing human labor with compute and that labor’s share of income may decline. If AI systems can replicate expertise at the cost of compute, employers may prefer to run additional instances of AI rather than hire human specialists — a shift that could reshape career incentives across fields.
Possible consequences for the discipline
If many early-career mathematicians choose to leave academia, the field could face a shortage of future researchers and instructors, making it harder to sustain and transmit the discipline’s traditions and methods. The survey does not predict the extinction of mathematics, but it does document rapid changes in attitudes and career plans that could affect how the field develops.
Conclusion
The MIT survey provides a quantitative snapshot of how AI’s recent progress is affecting the outlook of young mathematicians. The scale of the survey and the prevalence of pessimistic responses are noteworthy signals for universities, funders, and policymakers as they consider how to integrate AI into research, hiring, and education while addressing the career concerns of early-stage researchers.
