Mathematicians Split Over AI That Might Do the Math Better
AI·November 18, 2026
What does it mean to do mathematics when a machine might do it better? That question has moved from philosophy seminars into the daily work of mathematicians, and Scientific American senior reporter Joseph Howlett takes it on in the December/January cover story "The End of Math as We Know It." The piece is also the basis for the latest episode of the magazine's Science Quickly podcast, where Howlett reads it aloud.
The field is divided. Some of the leaders Howlett spoke with see AI as a way to finally make progress on old problems that have resisted human effort for generations. Others see it as a distraction, or a threat, to the core of the discipline, which many mathematicians value for the reasoning and understanding behind a proof as much as for the answer itself.
The episode lays out the choice in direct terms: should mathematicians welcome the technology as a boon to their work, or should they band together to resist it? That framing captures a community still working out how much of its identity depends on doing the work by hand.
The stakes reach beyond the research community. Mathematics underpins cryptography, engineering, physics and much of modern computing, so changes in how proofs are discovered and checked tend to spread into other fields. Whether AI speeds that process up or changes what counts as genuine understanding is still an open question, and the reporting suggests the answer will depend heavily on the choices mathematicians make in the next few years.