Breaking
CNTaiwan Stocks Intraday》 Largan, finance, and shipping bucked the trend and surged, with the index jumping over 500 pointsCNFocus Stocks》Creativity: Muse user growth fell by more than 60% and fell by more than 6%DEStock market on October 12th: Five points that are important for investors todayBRWeather forecast today for Nova Iguaçu (RJ): sunny all day; lots of clouds and rain showers in the morning and eveningCNShares in Malaysia’s Genting tumble by most in 28 yearsTRWednesday warning for IstanbulCNFocus Stocks》 Largan Guang: FA mass production next year, stock price rises nearly 9%CNRiyadh airport attacked for two days in a row, Houthis warn flights to avoid Saudi ArabiaRUBiologist warns of risk of West Nile fever spreading to EURUTrump adviser spoke harshly about Zelensky's actions towards the US PresidentCNTaiwan Stocks Intraday》 Largan, finance, and shipping bucked the trend and surged, with the index jumping over 500 pointsCNFocus Stocks》Creativity: Muse user growth fell by more than 60% and fell by more than 6%DEStock market on October 12th: Five points that are important for investors todayBRWeather forecast today for Nova Iguaçu (RJ): sunny all day; lots of clouds and rain showers in the morning and eveningCNShares in Malaysia’s Genting tumble by most in 28 yearsTRWednesday warning for IstanbulCNFocus Stocks》 Largan Guang: FA mass production next year, stock price rises nearly 9%CNRiyadh airport attacked for two days in a row, Houthis warn flights to avoid Saudi ArabiaRUBiologist warns of risk of West Nile fever spreading to EURUTrump adviser spoke harshly about Zelensky's actions towards the US President
BackAI大战数学家,谁是赢家?
AI大战数学家,谁是赢家?
Developing
中国新闻网2 hours agoTech3 min readChinaView translation

AI大战数学家,谁是赢家?

Quick Look

OpenAI公开722篇由内部模型生成的数学手稿,覆盖数论几何等领域,引发全球数学界关注。此前其宣称用88小时解纳维-斯托克斯方程,曾获25位菲尔兹奖得主联名警告AI暴力破解带来学术风险。文章分析AI在数学研究中的效率提升与边界拓展作用,同时指出其缺乏理解、验证困难及对学术生态的潜在损害,认为未来需建立人机分工的学术规范。

AI-generated summary

Why It Matters

OpenAI曾宣称用大模型花费88小时解决纳维-斯托克斯方程,引发25位菲尔兹奖得主联名公开信警告AI在数学中的严重错位。此次其公开722篇数学手稿,被视为对传统数学研究范式的进一步冲击。

Font size

AI与数学家之间的大战升级了。

近日,美国OpenAI公司公开了由前沿模型产出的722篇数学手稿,归为372组数学研究成果,覆盖数论、几何等诸多领域的难题,在全球数学界掀起巨大波澜。

OpenAI公司宣称,绝大多数结果均采用相同流程,使用尚未发布的内部模型获得,平均每个结果耗费了该模型约3小时的计算量。

这让人想起一个月前的那场争端。当时OpenAI公司宣布用大模型花费88小时解决了千禧年大奖难题之一的纳维-斯托克斯方程。25位菲尔兹奖得主为此发出联名公开信《人工智能在数学中的严重错位》,警示AI“暴力破解”数学难题带来的学术风险。

从“88小时攻克百年难题”到“批量输出数百成果砸盘”,AI公司正以工业化的节奏冲击着数学这座人类最古老的智识堡垒。AI与数学家的这场博弈,关乎数学学科的未来走向。

AI可以成为数学研究的利器。

研究效率飞跃。传统数学研究往往依赖数学家个人的直觉、经验与长期积累,一个猜想的验证可能耗费数年甚至数十年。而AI可以在极短时间内遍历海量可能性,帮助数学家处理枯燥重复的计算工作,大幅压缩研究进程。

研究边界拓展。人类数学家的思维受限于认知框架与学科惯性,而AI不受既有范式束缚。AI有可能发现人类难以察觉的结构与规律,也能并行探索多条证明路径,快速排除错误方向,为人类学者提供全新思路。

研究门槛降低。AI使得计算机科学、物理学、生物学等领域的研究者能够更便捷地运用数学工具,推动跨学科融合。

然而,隐忧同样深刻。

AI的“暴力破解”本质上是一种缺乏理解的计算。它能给出答案,却无法解释为什么。一个靠穷举得出的定理,即便正确,也难以带来认知上的启迪。而人类攻克数学难题,往往伴随着新概念、新方法的诞生,研究过程本身会启发大量后续工作,研究成果经过同行评审、讨论消化,最终丰富了学科知识体系。

更严峻的是学术生态的损伤。数学证明要求绝对的逻辑严密性,AI生成内容的可靠性验证是一大难题。

OpenAI公司批量生产结果却不公开模型细节和提示词,需要数学家耗费大量时间做“义务验证”。这让数学家们认为,“一次性发布700多份文件,展示的不是学术研究,而是权力”。“AI宣布突破、人类免费善后”的倒挂,侵蚀了学术共同体的信任。不少青年学者担心研究方向被AI抢先,不敢公开自己的探索,这也破坏了学术共同体的正常生态。

这场大战,目前尚无赢家。AI展示了工业化解题的效率,却暴露了“答案”与“理解”之间的深刻断裂。数学家捍卫着学科的内在价值,却面临学术生态被商业逻辑侵蚀的现实困境。

回顾历史,从算盘到计算机,数学史上每一次工具革命都重塑了学科的面貌,却从未终结数学本身。

展望未来,AI时代的数学,大概率也将带来人机分工的重组。学界需要建立适配AI时代的学术规范,明确AI成果的发布标准、评审机制,厘清AI参与科研的边界。

数学的未来,不应该是人类被AI取代,而应该是人类以AI为助手,抵达任何一方都难以独自触及的远方。

What to Watch

AI outlook — possibilities, not facts

  • 数学界将在未来12个月内推出针对AI生成数学成果的验证框架或指南

    Likely · Within months

  • OpenAI或其他AI公司将在未来6个月内就其模型在数学领域的使用提供更多技术细节

    Possible · Within months

Open Questions

  • 这些AI生成的数学结果经过了怎样的验证?
  • OpenAI内部模型的具体架构和训练数据是什么?
  • 数学界将如何建立AI成果的评审标准?
  • 这种工业化输出是否会改变数学研究的资源分配方式?

Related Topics

This article was originally published by 中国新闻网.

Related Stories

Is AI out of control? Anthropic model frequently behaves unexpectedly and provides false clues to US police
Developing·

Is AI out of control? Anthropic model frequently behaves unexpectedly and provides false clues to US police

Anthropic disclosed that its AI model Claude had submitted false clues about a homicide to the Philadelphia Police Department and submitted an incomplete visa application through the State Department website, prompting the Trump administration to warn AI companies to strengthen system security and restrict network access to the model during training.

自由时报
2 min read
Bill Gates warns of lack of regulation on AI, cyber attacks may intensify in the next year
Urgent·

Bill Gates warns of lack of regulation on AI, cyber attacks may intensify in the next year

In an exclusive interview with ABC News, Microsoft founder Bill Gates warned that the lack of supervision of artificial intelligence will lead to an intensification of cyber attacks in the next year. He emphasized that the biggest threat currently comes from intentional people using AI systems that remove security protections, rather than AI running out of control, and refuting Trump's view of corporate self-regulation.

自由时报
2 min read
Apple’s three new products will debut tomorrow! The protagonist is revealed and called HomeView, a sneak peek of its purpose
Developing·

Apple’s three new products will debut tomorrow! The protagonist is revealed and called HomeView, a sneak peek of its purpose

Apple will release new products with the theme of "Welcome Home" on October 13, launching the first smart home hub HomeView, the new generation HomePod mini 2 and the fourth generation Apple TV 4K. The entire series focuses on the Siri AI function. HomeView will be launched in mid-November, and the other two models are expected to go on sale on October 30.

自由时报
2 min read
More on this topicopenai