Updated
Updated · MIT Technology Review · Sep 22
OpenAI Faces Plagiarism Claims Over 2 Math Breakthroughs as Experts Challenge AI Hype
Updated
Updated · MIT Technology Review · Sep 22

OpenAI Faces Plagiarism Claims Over 2 Math Breakthroughs as Experts Challenge AI Hype

3 articles · Updated · MIT Technology Review · Sep 22

Summary

  • 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.

Insights

Are tech giants deliberately using superintelligence hype to hide massive security flaws and data theft from the public?
While we fear rogue AI, are data centers silently causing irreversible environmental damage under the guise of technological progress?
If AI only mimics understanding through statistics, what happens when critical infrastructure relies on a system that truly comprehends nothing?