菲尔兹奖得主联名警告AI数学风险
推荐指数 41.0 NO. 021 · 2026.09.12
发布2026/09/11Score143Comments8
为什么值得看
25位菲尔兹奖得主签署声明,指出LLMs数学能力突飞猛进的同时,其发展方向与数学界核心需求存在根本错位。这是顶级数学家首次集体对AI数学应用发出系统性警示,对依赖AI做定理证明或研究验证的团队有直接参考价值。
编辑判断
这份声明的签署速度本身就很说明问题——通常数学家群体以谨慎和共识驱动著称,但一周内从讨论到公开发布,说明内部紧迫感极强。值得注意的不是他们反对AI做数学,而是反对当前工业界主导的'解题竞赛'模式:LLMs被训练去生成看似正确的证明,却缺乏形式化验证所需的严谨结构。
这与Lean、Coq等交互式定理证明社区的理念形成张力。如果工业界继续用'解决了多少IMO题'作为数学AI的北极星指标,可能会挤占真正可验证、可审计的数学AI系统的资源投入。做AI4Math的团队需要重新评估自己的评估标准,特别是那些计划把LLM输出直接接入生产系统的场景。
社区反馈
意见分歧 8 条评论
核心争论:AI数学能力是替代人类还是辅助工具,旅程价值与结果价值之争
Comments moved to https://news.ycombinator.com/item?id=49662371.
In the Economist article Tao links, Hugo Duminil-Copin, draws a comparison: airdropping someone on the summit of Mount Everest is very different from climbing it. The fundamental issue with AI solving any perceived difficult problem is that we have lost the journey. The sight atop Mount Everest look
Sure, but actually what's the issue? You can still climb. If one do it for the sake of the journey, not some side quest to fame, glory, and social prestige, the mountain is still there. Maybe the one that lament that the sacred mountain should be left alone and be kept pristine of human arrogance an