Radio
Now Playing
Quickyla Radio โ€” Click to play
Open โ†’
3 min left
Back to News

NYU Mathematician Accuses OpenAI of Using Unpublished Data for GPT-4

NYU mathematician Alex Kim accuses OpenAI of using unpublished data to train GPT-4 for a purported solution to the Navier-Stokes problem. OpenAI denies the claims, citing public sources and ethical gโ€ฆ

OpenAI fought dirty on career-making math problem, says NYU mathematician
TechCrunch โ€” 8 September 2026
Text:
7 0 0

NYU mathematician Dr. Alex Kim claims that OpenAI engaged in unethical tactics to find a solution to the Navierโ€‘Stokes existence and smoothness problem, a question that carries a $1โ€ฏmillion bounty from the Clay Mathematics Institute. Kim alleges the company used proprietary data sets and undisclosed research to train its GPTโ€‘4 model, allowing it to generate a purported solution that it then submitted for the prize. The claim surfaced after OpenAI announced a partnership with the Institute to explore AIโ€‘assisted proof discovery, sparking controversy over the boundaries of academic integrity in machineโ€‘learning research.

The Navierโ€‘Stokes problem is one of the seven Millennium Prize Problems, and a solution would settle a fundamental question about the behavior of fluids in three dimensions. The Clay Mathematics Institute has offered a $1โ€ฏmillion reward for a rigorous proof since 1993, and the problem is central to fields ranging from aeronautics to climate science. For decades, mathematicians have tried to prove that solutions to the equations always exist and remain smooth over time, but no one has yet succeeded. A machineโ€‘generated proof would be a landmark achievement, but it also raises questions about how AI can contribute to mathematical discovery while respecting intellectual property and peer review.

Kim says that OpenAIโ€™s approach involved โ€œfighting dirtyโ€ by incorporating unpublished papers and private datasets that were not freely available. According to the mathematician, the companyโ€™s engineers used the bounty to pressure researchers into sharing confidential work, and the resulting AI model was then used to produce a solution that bypassed standard peerโ€‘review protocols. OpenAI has denied these allegations, stating that its training data come exclusively from public sources and that it follows strict ethical guidelines. The company has announced an internal review and said it will cooperate with any external investigation.

The case is now being investigated by the U.S. Federal Trade Commission, which is looking into potential violations of dataโ€‘usage laws. The Clay Mathematics Institute has also expressed concern and said it will review its partnership terms. If the allegations

Read Full Story at TechCrunch โ†’
Advertisement
React:
Sources
Sponsored

More to Read

Sophie Adenot completes second spacewalk, upgrades ISS commโ€ฆ
๐Ÿ”ฌ Science
Sophie Adenot completes second spacewalk, upgrades ISS communications antenna
France 24 ยท 13 days ago
See last nightโ€™s โ€˜blood moonโ€™ eclipse from around the world
๐Ÿ”ฌ Science
See last nightโ€™s โ€˜blood moonโ€™ eclipse from around the world
Scientific American ยท 11 days ago
How to see the 2026 โ€˜blood moonโ€™ eclipse this week
๐Ÿ”ฌ Science
How to see the 2026 โ€˜blood moonโ€™ eclipse this week
Scientific American ยท 14 days ago
Nigeria's jet fuel conundrum: Scarcity at home, abundance aโ€ฆ
๐ŸŒ World News
Nigeria's jet fuel conundrum: Scarcity at home, abundance abroad
DW World ยท 14 days ago
Firms scramble for battery power in Spain and Portugal
๐Ÿ’ฐ Business
Firms scramble for battery power in Spain and Portugal
BBC Business ยท 13 days ago
Is Sudanโ€™s battlefield shaping the terms of its next politiโ€ฆ
๐ŸŒ World News
Is Sudanโ€™s battlefield shaping the terms of its next political phase?
Al Jazeera ยท 13 days ago
Full view