Terence Tao warns OpenAI's bulk AI math proofs cause 'proof indigestion'
Mathematician Terence Tao and other scholars have criticized OpenAI for publishing 719 mathematics manuscripts on GitHub, said to contain solutions to hundreds of open problems. They argue the move conflicts with recommendations from a mathematics-and-AI advisory group on transparency and release norms for proprietary models, and has produced a kind of "proof indigestion": the AI delivers answers without human-readable explanations or scholarly exchange. The scholars also warn that treating problem-solving speed as the sole measure of mathematical progress is short-sighted and could crowd out traditional research paths and impoverish the field's creative soil.
The dispute puts the tension between AI's bulk output of mathematical results and the health of the research ecosystem on the table, potentially shaping norms for model releases and credit for results.
陶哲轩警示 OpenAI 批量生成证明,数学界面临“证明消化不良”

AIbase基地
发布于AI新闻资讯 · 1 分钟阅读 · Oct 9, 2026
10
著名 数学家陶哲轩等学者近期对 OpenAI 批量发布 AI 生成的数学证明提出严厉批评。事起 OpenAI 从 GitHub 发布了包含数百项开放难题解答的 719 份数学手稿,但此举被指违背了“数学与人工智能顾问小组”关于提高专有模型透明度与发布规范的建议。