Math group urges researchers to reject OpenAI over new AI papers

A group of mathematicians has asked researchers to stop working with OpenAI. The Association for Human Mathematics released this statement two days after OpenAI shared 722 math papers. These papers cover 372 groups of results and include solutions or progress on many hard problems in math and computer science.

The Criticism

The association criticised the AI firm for ignoring research norms and not giving proper credit. The group also brought up earlier complaints about illegal copying and trademark issues. Tensions grew when OpenAI claimed it solved the Navier-Stokes problem, which is one of the big Millennium Prize Problems. Critics said the company cares more about speed and AI outputs than open checks and independent reviews.

The Risks

The Advisory Group on Mathematics and Artificial Intelligence previously said big AI labs should not test hard math problems on internal models. The group warned this could create a two-tier system where labs run far ahead of human researchers. The math association called the release of over 700 files at once a show of power rather than real scholarship. Members urge the public and researchers to look at this publishing model with doubt and focus on human understanding.

Context

Artificial intelligence tools are being used more often in scientific fields, which has led to debates about ethics and credit. Professional bodies like the Association for Human Mathematics aim to keep research transparent and led by humans.