In the autumn of 2026, a quarrel among mathematicians over a claim made by OpenAI has opened a question older than any machine: who owns the work of understanding, and how far can a result be believed before people who know the terrain have tested it. The company says its AI overturned a thirty-year-old assumption in computer science, the Exponential Time Hypothesis, and the field is divided over whether that claim stands and whether the credit has been rightly placed. Beyond the single hypothesis, the unease reaches toward a larger fear that science may be slowly blurred by the authorship of
Mathematicians Split Over OpenAI's Claims That AI Can Refute Long-Held Computer Science Assumptions
A bust-up between mathematicians has become a question about trust in science
Mark
Start with the basics. What is actually being claimed here?
Mimi
OpenAI says its AI disproved the Exponential Time Hypothesis, an assumption in computer science that has stood for about 30 years.
Mark
And mathematicians disagree about that?
Mimi
They're publicly split. The coverage calls it a bust-up, and it has spread well beyond specialist circles.
Luke
Before we go further: what do we actually know about the split? The headlines say divided, but I can't tell you who is on which side or what their objections are.
Mimi
That's fair. The material doesn't spell out the technical arguments. We know the dispute exists, not its detailed shape.
Mark
Why does credit come up at all?
Mimi
Critics are questioning whether credit for the mathematical work is fairly attributed. Scientific American's wording, that OpenAI scooped mathematicians' work, points at the same tension.
Luke
Careful, though. That's a headline's characterization, and a questioning of credit isn't a finding that credit was misassigned.
Mark
So what's the bigger worry?
Mimi
The BBC frames it as fear that AI could undermine science, if claims outrun the checking.
Luke
And whether the result is correct is still open until independent experts weigh in. That's the thing to watch.
Le Pouls
- Mathematicians are openly split over whether OpenAI's AI truly refuted a three-decades-old assumption in computer science.
- The Exponential Time Hypothesis claim has become a flashpoint, pulling a specialist disagreement into mainstream headlines.
- Critics are pressing hard on who deserves credit, with talk of OpenAI having scooped the mathematicians' own work.
- Fears are spreading that AI-generated claims could erode trust in science before experts have had a chance to check them.
- Everything now waits on independent verification, and the outcome will shape how AI-driven research is credited and trusted.
In the autumn of 2026, a quarrel among mathematicians over a claim made by OpenAI has opened a question older than any machine: who owns the work of understanding, and how far can a result be believed before people who know the terrain have tested it. The company says its AI overturned a thirty-year-old assumption in computer science, the Exponential Time Hypothesis, and the field is divided over whether that claim stands and whether the credit has been rightly placed. Beyond the single hypothesis, the unease reaches toward a larger fear that science may be slowly blurred by the authorship of machines, and the answer will depend less on any headline than on whether independent minds can confirm what was said.
The story arrives as a scatter of headlines, and their differing wording shows how quickly the argument has spread. The BBC frames it as a bust-up between mathematicians that feeds fears AI could undermine science. Scientific American says OpenAI scooped mathematicians' work on another million-dollar problem, while New Scientist reports that AI has refuted a 30-year-old assumption in computer science. Business Insider has gathered what experts say about the claim.
At the heart of it is the Exponential Time Hypothesis, a long-standing assumption that OpenAI says its AI has disproved. Mathematicians are publicly divided over that claim. The available material does not show which side each camp takes or lay out their technical reasoning, but it does confirm that the disagreement has moved out of specialist circles and into the mainstream.
Two questions appear tangled together: whether the result holds, and who should be credited for the mathematics behind it. Critics doubt the credit is being fairly shared, and the hint of a scooped team suggests this is not the first time such comparisons have arisen. OpenAI points to its work with academic communities, though the sources do not say how it has answered the specific objections.
What comes next is verification. Independent experts confirming or rejecting the result will matter most, and how this dispute settles may shape how AI-generated research is credited and trusted in the years ahead.
The story reaches us as a cluster of headlines rather than a single article, and the headlines themselves show how far the argument has traveled. The BBC describes a bust-up between mathematicians that is fuelling fears AI could undermine science. Scientific American says OpenAI scooped mathematicians' work on another million-dollar problem. New Scientist reports that AI has refuted a 30-year-old mathematical assumption in computer science. Business Insider has collected what smart people are saying about OpenAI's Exponential Time Hypothesis claim.
At the center is that claim: OpenAI says its AI disproved a long-standing assumption in computer science, one the field has worked under for roughly three decades. The assumption is the Exponential Time Hypothesis, and the claim to have refuted it is what has set off the dispute.
Mathematicians are publicly divided over it. The material available here does not lay out who stands on which side, and it does not give the technical arguments each camp is making. What it does establish is that the disagreement is open and that it has moved beyond a quiet exchange among specialists into mainstream coverage.
Two questions seem to be tangled together. One is whether the result holds up. The other is who deserves credit for the mathematical work behind it. Critics are questioning whether that credit is being fairly attributed, and the Scientific American headline, with its language of OpenAI having scooped mathematicians' work, points the same way. The phrase about another million-dollar problem suggests this is not the first time the company's work has drawn this kind of comparison.
The BBC framing widens the stakes. The worry it names is not only about one hypothesis but about whether AI-generated claims could undermine science more broadly, by blurring who did the work and how far a result can be trusted before experts have checked it.
OpenAI, for its part, has a page describing its work with academic communities to build scientific understanding, which appeared alongside the coverage. The source material does not say how the company has responded to the specific criticism.
What to watch is verification. Whether independent experts confirm the result will matter first, and the way the dispute resolves could shape how AI-generated research is credited and trusted afterward.
Citations marquantes
The dispute is fuelling fears that AI could undermine science.— BBC headline, paraphrased
Envie de l'histoire complète ?
Lire l'original sur Google News ↗
Technology