Responsible Release of AI-Generated Mathematics

# Summary Frontier AI labs are testing advanced mathematical problems on proprietary models while the broader scientific community lacks access, prompting concerns about maintaining mathematical integrity and human understanding. The document recommends that AI labs either release mathematical results accompanied by human understanding following traditional academic norms, or if results remain ununderstood, provide substantial funding and support to help the mathematical community develop that understanding organically. The authors emphasize that human understanding of mathematics must remain paramount and community-led, not directed by the AI labs that produced the results.