Paper: training mathematicians in the AI era — agency and cognitive offloading
ceciletamura · x · 2026-10-11
Thomas Koberda of the University of Virginia published a 17-page paper, "Training Mathematicians in the Age of AI: Intellectual Agency, Cognitive Offloading, and the PhD Thesis."
Core argument: powerful AI is weakening the traditional link between mathematical output and evidence of mathematical expertise — producing an original theorem or a polished dissertation can no longer by itself certify the intellectual formation of its nominal author. Koberda argues graduate math education should be organized around forming intellectual agency: internal technical competence, mathematical judgment, understanding and responsible participation in a shared intellectual culture.
He distinguishes productive from premature cognitive offloading, proposes complementary independent and AI-augmented training modes, and suggests reformulating the role of the PhD thesis and defense. More broadly, he argues academic mathematics should see itself increasingly as an institution for reproducing and stewarding human mathematical expertise rather than primarily as a theorem-producing mechanism.
Related event: UVA Professor Discusses Training Mathematicians in the AI Era(2 posts)→
More from AGI Musings
- One unilateral OpenAI decision could devastate an entire field of math research — DavideCrapis · 2026-10-11
- Wes Pegden: in 25 years AI could make decisions we no longer comprehend at all — littmath · 2026-10-11
- Your Body Runs on Just 80.6W Across 37 Trillion Cells, Argues 'Natural Intelligence' Talk — vinodg · 2026-10-11
- AGI Has Arrived? Why Nobody Can Agree on What AGI Actually Means — annetgriffin · 2026-10-11
- AI agents overstate results, far from autonomous research: Epoch AI and Anthropic studies — The Decoder · 2026-10-11
- At AGI's edge, people split into pleasure-seekers vs meaning-makers — danfaggella · 2026-10-11