Mathematician predicts AI will transform math research
8 reported
A mathematician writing to the Marginal Revolution blog described how artificial intelligence tools are changing mathematical research. The author stated that researchers can now test computational guesses with minimal effort and run incomplete drafts past AI systems like Claude and ChatGPT. The mathematician noted that AI is particularly effective on topics slightly off the beaten track, where basic techniques such as graph theory, algebra, and calculus can be applied. The author also observed that younger researchers are developing a better feel for different AI models by earning pocket money through AI training. The mathematician predicted that refereeing will change soon, as running published papers through AI as a control could reveal errors. The author concluded that the coming changes in math will be larger than the revolution in the early 1990s following work by Witten and string theory predictions about 4-manifolds.
What’s reported
The author is a mathematician who wrote to Marginal Revolution about AI in math research.
Researchers can now test computational guesses with minimal effort using AI.
Incomplete drafts and proofs of special cases can be run past Claude and ChatGPT.
The author stated they are not as good at linear algebra as AI systems.
AI is most effective on topics slightly off the beaten track, not esoteric subjects.
Younger researchers are making pocket money doing AI training and developing a feel for different AI models.
The author predicted refereeing will change, with published papers run through AI as a control.
The author said the coming changes will be larger than the early 1990s math revolution.
Key figures
GA (author, mathematician, not further identified in the article)
Sources: marginalrevolution.com
