.png){kind=logo}
OpenAI has announced three new scientific discoveries spanning mathematics, computer science, and physics, and is adopting a new approach to sharing AI-generated research with the academic community.
The announcement comes weeks after the company released more than 700 pages of solutions and progress on 372 major mathematical problems. This time, OpenAI has introduced a process in which external experts examine, explain, and evaluate the findings.
According to the company, the latest results cover harmonic analysis, computational complexity and quantum field theory. While two research manuscripts are already available publicly, the physics findings will be released through a separate collaborative process.
OpenAI has used three different strategies to share its latest discoveries. For mathematics, the company has published the AI-generated output, allowing researchers to study the findings and develop their own explanations.
For computer science, OpenAI has formed an external group of collaborators to explain the results through their own write-ups and supporting materials. The company has also made the AI output publicly available for independent examination.
The physics research will follow a different route, with OpenAI deferring the release of the materials to the expert group examining the findings. The company has also invited feedback on how to broaden participation and make the process more useful.
Also Read: OpenAI GPT-6.1 Terra Launches with App Control and Faster AI Workflows
One of the two newly published manuscripts establishes Sogge's local smoothing conjecture for waves in every spatial dimension. The conjecture concerns the behavior of solutions to the wave equation, a mathematical framework used to describe wave propagation.
OpenAI's models produced a proof of the local smoothing conjecture across all dimensions. The company says the result also resolves the Kakeya conjecture and Stein's restriction conjecture in all dimensions.
The company has made the mathematical output available for researchers to examine and interpret independently.
The second manuscript presents a proof claiming that the Exponential Time Hypothesis (ETH) is false. The hypothesis is widely believed to be true and assumes that certain computational problems cannot be solved efficiently.
ETH has served as a foundation for hundreds of research papers across fields including databases, computational biology, and computational geometry. If the new result withstands scrutiny, it could inform the development of more efficient algorithms.
Researchers associated with the original hypothesis are part of the external group examining and explaining the findings.
The third discovery concerns interacting quantum field theories with gauge fields and fermions in four dimensions. The work includes quantum chromodynamics (QCD), which describes the strong interaction between quarks and gluons.
The external group reviewing the physics findings has yet to publish its assessment. OpenAI has left the release of the research materials to this group.
The company has also said it will not seek academic credit for discoveries made by its models. Instead, the participating researchers will receive credit for the explanations and materials they produce, while ensuring that relevant researchers, techniques, and ideas are appropriately acknowledged.
