OpenAI publishes 722 maths papers, but mathematicians aren't convinced
OpenAI released 722 papers. These papers claim to solve famous maths problems. Experts say many are flawed. Many have already been retracted.
OpenAI this week published hundreds of maths papers. OpenAI claims these papers could advance human knowledge.
An unreleased AI model made 722 papers. The 722 papers address 372 open problems in algebra, geometry and computer science.
Some papers claim breakthroughs on high-profile problems like the Riemann hypothesis. Each is worth US$1 million if solved.
The release has sent shockwaves through the maths community. Several papers have already been retracted or amended.
One mathematician told the AI something important. At least one result was a 'serious hallucination'.
It should never be cited without expert checking. Others say the papers are poorly written.
They would be rejected instantly by academic journals. OpenAI says it aims to enable progress.
But the scale and poor quality have created confusion instead.
- 722
- Papers published
- 372
- Open problems addressed
- $1 million
- Riemann hypothesis bounty
- Millennium Prize Problems
- Prize problems claimed
Why it mattersIt shows how AI can make answers that sound right but fall apart when experts check them. For mathematicians, it raises questions about how to handle so much AI-made work.
AustraliaNo direct Australian effect yet, but as AI becomes more used in research globally, Australian scientists will face the same challenge of sorting credible AI results from confident-sounding errors.
✓ Claims checked against the source and corrected before publish. checked 2 h ago
Open this story in InSnip →
