OpenAI AI Models Prove Long-Standing Math Theorem

OpenAI announced on Monday that its artificial intelligence models have identified three new scientific breakthroughs in mathematics, computer science, and physics. For mathematicians and physicists, this marks a shift where algorithms are not just assisting with calculations but producing original proofs in core theoretical fields.

The first discovery involves harmonic analysis, a branch of mathematics that breaks down complex signals into elementary waves. OpenAI states that its models proved the local smoothing conjecture in every spatial dimension. Previously, researchers had only managed to prove the theory in three dimensions.

Alex Iosevich, a mathematics professor at the University of Rochester, described the event as a shock. He noted that the conjecture was formulated by his doctoral advisor, Chris Sogge, making the result deeply personal. Iosevich called it a historic development and said he feels joy and optimism about the future of scientific research.

In the field of complexity theory, the AI models disproved the Exponential Time Hypothesis. This hypothesis relates to the time required to solve complex computational problems. Disproving it changes the understanding of how efficiently certain algorithms can run.

The third discovery concerns quantum field theory. OpenAI claims its models provided mathematically rigorous proof of the existence of quantum chromodynamics. This field studies the interactions between quarks and gluons, the fundamental particles that make up protons and neutrons.

The company is working with outside scientists to turn the raw research data into scientific understanding. OpenAI selected three expert groups to act as shepherds for the findings. The groups handling mathematics and computer science chose to release the data publicly immediately. The physics group decided to delay the release to write their own interpretation of the results first.

OpenAI acknowledged that these different release strategies involve tradeoffs. The company, valued at $852 billion, stated that empowering experts helps contextualize the results for the academic community. However, choosing specific shepherds creates an asymmetry of information regarding who gets involved first. The company said it welcomes feedback on how to broaden participation in this process.

This announcement follows a similar claim by Anthropic last month. That company said its AI model, Claude, discovered a new enzyme resembling CRISPR. Scientists reacted to the Anthropic finding with mixed opinions, with some calling it exciting and others arguing the company exaggerated its significance. OpenAI’s release adds to the growing list of AI-credited scientific advances.