Mathematicians' response to OpenAI's proofs turned on Tuesday after OpenAI released 722 maths papers across fields from algebraic geometry to number theory. The release arrived less than a month after OpenAI drew attention for its solution to the long-standing Navier-Stokes problem, and the reaction was immediate.
The scale was what unsettled the field. OpenAI said it had consulted extensively with mathematicians before publishing, yet mathematicians and commentators still described the batch as shocking, heartbreaking and a display of power rather than scholarship.
Scott Aaronson and the shock
Scott Aaronson, chair of computer science at the University of Texas at Austin, compared the moment to a hunter-gatherer who has spent his entire life learning to survive deep in an unforgiving rainforest and then sees a giant resort hotel with a helipad and heated pools and Airbnbs appear beside him. The image captured the sense that the pace and scale of the release had changed the ground under mathematicians who expected a slower, more human pace of discovery.
Jordana Cepelewicz, maths editor at Quanta Magazine, called the development heartbreaking for the mathematics community. She said problems in mathematics are posed because they represent interesting journeys, while AI models just skip to the end.
The floodgates are open
Joseph Howlett, writing in Scientific American, said the floodgates are open and called the release the biggest maths breakthrough in two decades. The Association for Human Mathematics took the opposite view, saying that releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power.
In the same statement, the Association for Human Mathematics urged mathematicians to discontinue their work with OpenAI and to return to a vision of science that centres human understanding. That warning reflected a wider fear in the field that new blood will no longer be attracted to a discipline being remade at speed.
Publishing habits and cryptography
Researchers have already started changing how they write papers. Some have stopped positing open conjectures at the end of their work so they are not scraped by bots, and others are publishing prematurely to avoid getting scooped by AI.
Anthony Cuthbertson said cryptography forms the backbone of online encryption used in web browsers, secure messaging and online banking, and that the mathematical problems behind it are currently too difficult to solve efficiently. He said there is no evidence that AI models have come close to breaking those problems, but the latest advances could lead to new algorithms capable of cracking them.
The consequence for mathematicians is not simply that OpenAI has published more work. It is that the speed, size and claimed reach of the release have changed what counts as a safe pace for publication, and the field now has to decide whether to adapt its habits or pull back from OpenAI altogether.
