Emerging evidence
More
Math students fear AI will take their jobs-they’re using it anyway
- Evidence type
- SurveyPreprint, not yet peer reviewed
- Sample
- n = 268
- Sources
- 1 sourceJournalism
- Published
- 2 min read
- Truvace record
- TRV-2026-2989cite or check it
A survey of 268 graduate students and postdocs reveals researchers are unsure of their future in mathematics and uneasy about using AI—the very technology they fear Everyone has heard what the giants of mathematics think about artificial intelligence—the open letters, the op-eds, the predictions from laureates who have already cemented their reputation. Almost no one has asked the people whose future is less sec

The problem
In a voluntary survey of 268 math graduate students and postdocs, seven in 10 said they were less likely to remain in math because of AI, and two-thirds said they felt less motivated to do their research.
In brief
A survey of 268 math graduate students and postdocs from nearly 60 institutions found that seven in 10 are now less likely to remain in math, and two-thirds feel less motivated to work on their research because an AI model might beat them to their goals. The survey was run by three math Ph.D. students and is described in a preprint.
Respondents broadly expect AI to make tenure-track jobs harder to get. Woodruff, one of the organizers, said some people told her they felt useless, and that for many mathematicians the field is part of their identity. Early-career researchers worry about the next one or two years, not 20, and doubt they will have jobs or funding.
Views on the technology itself are split. About half are optimistic about AI's potential for mathematics and a third oppose its use, either outright or for creative work. Nearly all respondents use AI anyway, and many wrote that they felt pressured to keep pace.
Main points
- The online Google Forms survey drew 268 responses from nearly 60 institutions and spread by word of mouth and department mailing lists.
- Seven in 10 respondents said they are less likely to remain in math, and there was broad agreement that AI will make tenure-track jobs harder to get.
- Two-thirds reported feeling less motivated to work on their research.
- About half are optimistic about AI's potential for math, while a third oppose its use, either outright or for creative mathematical work.
- Among the optimists, four in five said they are excited about AI in theory but strongly opposed to how things are unfolding in practice.
- Just 13 percent of all respondents are happy with the current trajectory.
- Nearly all respondents use AI, but many wrote that they feel pressured to use it to keep pace.
The full story
Three math Ph.D. students, Anna Brandenberger, Hanna Mularczyk and Dora Woodruff, launched the survey after AI cracked problems that mathematicians had chased for decades. They wanted the voices of graduate students and postdocs alongside those of senior figures who have issued open letters and predictions.
The results show a community that is frustrated and anxious yet divided on what to do. Woodruff said some respondents feel that their work, minds and creativity do not matter, and that math is tied to many people's identities.
Attitudes toward AI are mixed even among its users. Optimists mostly object to current practice, only 13 percent are happy with the present trajectory, and many described reluctantly mastering a technology they fear could make them obsolete.
Sources
1 on file- Scientific Americanscientificamerican.com
Talk it over
Seven in 10 of the 268 math graduate students and postdocs said they are less likely to remain in math, yet nearly all use AI. Why might people use a tool they fear?
Math students fear AI will take their jobs-they’re using it anywayhttps://truvace.com/item/math-students-fear-ai-will-take-their-jobs-theyre-using-it-anyway
Seven in 10 of the 268 math graduate students and postdocs said they are less likely to remain in math, yet nearly all use AI. Why might people use a tool they fear?
Answer in the threadThe survey was voluntary and spread by word of mouth and department mailing lists. How might that have shaped who answered, and which young mathematicians may be missing?
Answer in the threadNearly all responses arrived before OpenAI published AI-assisted solutions to hundreds of open problems in October. What would a repeat survey need to show to tell you the fears have grown or eased?
Answer in the thread
If you have more time
Woodruff said early-career researchers worry about the next one or two years, not 20. What could departments or funders do in that window, and who should decide?
Answer in the threadAmong the optimists, four in five said they oppose how things are unfolding in practice. What would you ask them to learn what practice they object to?
Answer in the thread
ace

What readers say