Artificial intelligence has achieved something once thought impossible: helping solve part of a mathematical problem that has puzzled researchers since 1946.
An OpenAI reasoning model discovered a new approach to the famous Erdős unit distance problem, challenging assumptions that mathematicians had accepted for decades. Human researchers subsequently verified the result.
The achievement is significant not simply because of the mathematics involved, but because it suggests AI may increasingly contribute to original scientific discovery rather than simply processing existing knowledge.
