OpenAI announced on Monday that its AI models have proved the local smoothing conjecture in all spatial dimensions for the first time, a result that had previously only been demonstrated in three dimensions. The conjecture, a central proposition in harmonic analysis, concerns how complex signals and functions break down into elementary waves.
Alex Iosevich, a mathematics professor at the University of Rochester who specializes in harmonic analysis, told Al Jazeera that OpenAI appeared to have "cracked" the conjecture in a "historic development." Iosevich noted that the conjecture was formulated by his PhD adviser, Chris Sogge. "The news was a shock at first," Iosevich said. "Now I am filled with joy and optimism, because the scientific enterprise as a whole is going to make giant leaps as a result of the AI revolution."
Iosevich expressed mixed emotions about the shift. "I feel a twinge of sadness as a result," he said, adding that many theoretical mathematicians share the feeling. "But I firmly believe that these feelings will soon give way to the realisation that there is so much we can now accomplish."
OpenAI's models also disproved the Exponential Time Hypothesis, which relates to solving complex computational problems, and provided what the company characterized as "mathematically rigorous" proof of the existence of quantum chromodynamics, which studies the interaction of quarks and gluons.
OpenAI said it consulted with scientists outside the company to convert its findings into scientific understanding. The company worked with three expert groups acting as "shepherds" to determine release strategies. The groups collaborating on mathematics and computer science results chose to release the AI-generated data publicly immediately. The physics group opted to defer release while writing its own interpretation of the findings.
OpenAI acknowledged tradeoffs in its approach. "The collaborations empower experts to unpack the results and contextualize them in the broader academic community," the company stated. "On the other hand, choosing initial shepherds necessarily shapes who gets involved first and creates asymmetry of information."
This announcement follows similar claims from other AI companies. Last month, Anthropic announced the discovery of a new enzyme resembling CRISPR, the gene editing mechanism used to treat cancer and sickle cell disease. That announcement drew mixed reactions, with some scientists calling it exciting while others suggested the company had overstated its significance.