OpenAI Launches AI Research Breakthroughs in Three Fields, Why Is the Academic Community Not Buying It?

On Monday, OpenAI announced in an official statement that it has made breakthrough progress in three fields: mathematics, computer science, and physics, and will adjust the way it communicates the release of research results to the public.

While this seems to be a correction of its previous approach, researchers interviewed by The Verge are growing increasingly impatient with OpenAI's actions.

Over the past few months, OpenAI's relationship with the mathematics community has been less than pleasant. In early October, the company released hundreds of unreviewed preprints related to AI's mathematical capabilities all at once, pushing them directly to researchers without any prior notice. Preprints are initial research drafts that have not yet undergone peer review and have not been formally published; such content is not typically released in bulk in standard academic communications. Researchers have to invest a lot of extra time verifying the authenticity of conclusions and checking for derivation errors, and the chaos caused has not yet been fully resolved. The Verge also did a special report on this matter at that time.

The new plan OpenAI has proposed is to collaborate with expert "shepherds" in the field, who will determine the most appropriate way and pace to disclose new results. Details of the three specific research results have not yet been fully released, and the full content can be viewed in the original The Verge report.