OpenAI AI math feat stirs plagiarism dispute

OpenAI's use of artificial intelligence to solve a major mathematics problem has intensified debate over the future role of human mathematicians and prompted a plagiarism dispute.

The company used thousands of AI agents this month to solve the Navier-Stokes equations, a century-old challenge, in four days. Terence Tao, a UCLA professor and Fields Medal winner, said he was “completely overwhelmed with an existential crisis” in his field.

The development follows several AI-linked mathematical advances this year. Antonio Auffinger of Northwestern University said AI systems had improved from frequent hallucinations to producing long chains of reasoning without losing the thread.

But some mathematicians say OpenAI's work relied heavily on years of human research. Benjamin Antieau of Northwestern University said there was “awe and amazement” around the result, but that the system was following approaches developed by other mathematicians over many years.

Tristan Buckmaster and Levent Alpoge say they had come close to the solution before OpenAI's result and are questioning whether their work was plagiarized. OpenAI denies that charge.

Tao was among 25 Fields Medal laureates who signed a letter last Friday warning of a potential “destructive effect” from AI on mathematics. Stanford professor Mohammed Abouzaid said AI systems still had not shown the kind of creative leaps involved in creating new mathematics.

Anima Anandkumar of Caltech said her team solved a simplified form of the equations using a physics-informed neural network that did not rely on prior researchers' work, though she said the model still had to be designed and tested by people.

Auffinger said he did not expect AI tools to answer every mathematical question and said human mathematicians would still have an important role in deciding which questions matter.