OpenAI models verify 30-year-old math conjecture and disprove complexity theory assumption

OpenAI announced on Monday that its artificial intelligence models generated three distinct scientific results in mathematics, computer science, and physics.

The company stated that its systems proved the local smoothing conjecture in every spatial dimension. Researchers in harmonic analysis, a field that breaks down complex signals into elementary waves, had previously only verified the theory in three dimensions.

Alex Iosevich, a mathematics professor at the University of Rochester, described the result as a historic development. He noted that the conjecture was formulated by his PhD adviser, Chris Sogge. Iosevich said the news was a shock at first but now fills him with joy and optimism about the scientific enterprise.

He added that he feels a twinge of sadness regarding the changing nature of theoretical mathematics, though he believes this feeling will give way to the realization of new accomplishments. Iosevich emphasized that many pure math problems originated from physical considerations and can now be explored rapidly.

The announcement included the disproof of the Exponential Time Hypothesis. This hypothesis relates to the time required to solve complex computational problems. OpenAI also provided mathematically rigorous proof of the existence of quantum chromodynamics, a theory describing the interaction of quarks and gluons.

The San Francisco-based company said it contacted scientists outside the firm to help convert raw research data into scientific understanding. It consulted with three expert groups, which it called shepherds, to determine how to release the findings.

The groups responsible for the mathematics and computer science results chose to make the AI-produced data public immediately. The physics group decided to defer the release to write their own interpretation of the findings first.

OpenAI acknowledged that these approaches involve tradeoffs. The company said the collaborations empower experts to unpack the results and contextualize them for the academic community. However, choosing initial shepherds shapes who gets involved first and creates an asymmetry of information. The company said it welcomes feedback on how to broaden participation in this process.

This announcement follows a series of claimed scientific breakthroughs credited to AI. Last month, Anthropic, the creator of Claude, announced the discovery of a new enzyme. That enzyme reportedly resembles CRISPR, a gene editing mechanism used in treating conditions such as cancer and sickle cell disease. Scientists reacted mixedly to the Anthropic announcement, with some calling it an exciting development and others suggesting the company exaggerated its importance.