Amateur mathematician Liam Price, 23, solved a 60-year-old math problem using ChatGPT Pro, a method no human had previously considered. The problem, related to primitive sets of numbers, had stumped experts for decades. Price, with no advanced math training, used the AI’s GPT-5.4 Pro subscription to generate a solution posted on www.erdosproblems.com. This marked a rare instance of AI contributing original insights to complex mathematical conjectures.

The problem involves Erdős sums, scores calculated for primitive sets. Terence Tao, a UCLA mathematician, noted the AI’s solution avoided conventional approaches, using a formula from related math fields. Experts like Jared Lichtman from Stanford highlighted the method’s novelty, suggesting it could unify disparate problems. The proof required refinement but revealed a new perspective on number theory.

Price’s work, initially a casual experiment, underscores AI’s potential in mathematics. While prior AI solutions lacked originality, this case shows how LLMs might bypass human cognitive blocks. The new method could advance number theory and inspire further AI applications in academia.

This breakthrough highlights the evolving role of AI in solving longstanding problems. It raises questions about how machine-generated insights might reshape mathematical research and education.