Mathematicians have accused OpenAI of plagiarism and research misconduct after the company touted Astra as solving decade-old problems, then experts concluded the results were less novel than first presented.
Two days before OpenAI announced another math breakthrough, NYU professor Tristan Buckmaster published a statement alleging the company had taken others’ work and misattributed it.
The report argues those claims fit a broader pattern in which AI companies frame products as near-superintelligent while critics say the real issues are weak scientific practice, poor security, and lack of accountability.
Hundreds of mathematicians have urged policymakers to rely on independent experts rather than company press releases, warning that commercial incentives are driving exaggerated capability claims.