OpenAI releases hundreds of AI-generated mathematical results, prompting expert concern over verification and field impact

OpenAI abruptly released nearly 400 AI-generated mathematical results distributed across more than 700 manuscripts, spanning disciplines such as combinatorics, geometry, number theory, theoretical computer science, algebra, topology, probability, statistical mechanics, and mathematical physics. The scale of the release was described by mathematicians as 'staggering,' 'overwhelming,' and 'unprecedented,' with researchers noting that simply digesting the breadth of the work could take years.

Mathematicians interviewed by The Verge characterized the deluge as 'pure insanity,' expressing a mix of awe, excitement, and deep anxiety about its implications. Álvaro Lozano-Robledo, a professor of mathematics at the University of Connecticut, told The Verge that even reviewing the roughly 40-page table of contents and abstracts took him the better part of an hour, underscoring the overwhelming volume of material.

The release included formalizations in Lean, a programming language and proof assistant used to computationally verify mathematical proofs. However, OpenAI acknowledged that the results are 'at different stages of verification,' with fewer than half of the manuscripts described formally as of the time of reporting. OpenAI stated that only 300 top-line results out of 719 manuscripts had been formalized, representing approximately 42 percent of the collection.

Researchers raised concerns about the consistency and quality of the formalizations, noting that verifying even formalized proofs is a time-consuming process. Kevin Buzzard, a mathematics professor at Imperial College London, reported that in his field of algebraic number theory, only about six theorems 'stood out' and few appeared to be formally verified in Lean. He emphasized the difficulty of distinguishing valid results from 'slop'—a term used to describe low-quality, often erroneous AI-generated material that has proliferated in academic contexts.

The release has reignited debates about AI's role in academic research, particularly regarding attribution and rigor. OpenAI's previous mathematical outputs were criticized for poor or nonexistent attribution, and researchers told The Verge they had anticipated the flood of papers with terms like 'slopocalypse.' The lack of consistent formalization and verification has amplified these concerns, with experts warning that the sheer volume of material could overwhelm the field's capacity to assess and integrate new findings.