LectureRechercheRadarCadre d'investissement
Connexion / Inscription
← Retour à la lecture
SujetRédaction DisconfirmAIMis à jour 2026-11-09 01:54

伯克利AI批改争议

4 étapes2 pièces4 Sources

加州大学伯克利分校数学教授、菲尔兹奖得主Richard Borcherds宣布停止按正确性批改学生作业,转而仅检查是否完整提交。这一决定源于AI能力的飞跃:OpenAI近期宣称其模型已解决包括Navier-Stokes存在性与光滑性问题在内的多个数学难题,使得区分人类与AI生成的作业变得几乎不可能。Borcherds认为,继续按正确性评分不仅浪费资源,还变相惩罚诚实的学生并鼓励作弊。目前,他主要依靠禁止电子设备使用的线下考试来评估学生能力。与此同时,伯克利校方测试了Turnitin等AI检测工具后认定其不可靠,遂将制定课程AI政策的权力下放给教师。这一事件引发了关于数学专业未来价值的激烈讨论,部分学者担忧AI可能取代人类数学家,而另一些则呼吁建立新的教育“社会契约”。

Comment on en est arrivé là

  1. 2026-10-25

    《纽约邮报》报道,Richard Borcherds因发现学生普遍使用AI完成作业并获得高分,决定停止按正确性批改作业。

    nypost.com ↗
  2. 2026-10-26

    Borcherds进一步透露,由于在线考试平均分异常升高且无法有效识别AI作弊,他已将评估重心转向禁止电子设备的线下考试。

    nypost.com ↗
  3. 2026-10-29

    德国科技媒体t3n.de报道,Borcherds指出OpenAI已解决Navier-Stokes问题等数学难题,因此按正确性评分已无意义。

    t3n.de ↗
  4. 2026-11-08

    t3n.de更新报道,提及伯克利大学测试Turnitin和Rumi等AI检测工具失败,认定其不适合标准化部署,导致教师需自行制定政策。

    t3n.de ↗

Ce qui reste non résolu

  • 伯克利大学或其他高校是否会出台统一的AI作业评估标准,还是继续维持教师自主决定的现状?
  • 随着AI在数学领域能力的进一步提升,数学专业的招生规模和课程设置将如何调整以应对潜在的职业替代风险?
Connexion / Inscription
LectureRechercheRadarCadre d'investissement
  • 是否有更可靠的AI内容检测方法或评估模式(如口试、过程性评估)被证明能有效平衡学术诚信与学习激励?
  • Sources

    • nypost.comNew York Post: UC Berkeley professor refuses to grade students’ homework after spotting surge in AI usage↗
    • nypost.comNew York Post: UC Berkeley prof who quit grading student’s homework deploys new tactic to catch out cheaters↗
    • t3n.det3n.de: Mathe-Professor kapituliert: Warum Hausaufgaben durch KI sinnlos sind↗
    • newsweek.comNewsweek: Berkeley professor stops grading homework for accuracy over concerns that AI is making it harder for educators↗

    Couverture dans ce sujet

    • Richard Borcherds, Mathematics Professor at the University of California, Berkeley: AI Renders Homework Grading Meaningless2026-11-08Actualités

      Mathematics professor Richard Borcherds of the University of California, Berkeley, has publicly stated that grading mathematics homework is no longer meaningful due to students' widespread use of AI and its rapidly improving capabilities. He currently only checks for complete submission and admits uncertainty about whether he should still encourage students to major in mathematics.

    • UC Berkeley Math Professor Richard Borcherds: Grading for Accuracy Is Meaningless After AI Can Solve Problems2026-10-29Actualités

      Richard Borcherds, a mathematics professor at the University of California, Berkeley, argues that grading student assignments for accuracy has lost its meaning now that AI can solve complex mathematical problems, including the Navier-Stokes existence and smoothness problem. He currently grades only on whether assignments are submitted in full, while admitting uncertainty about whether he should still encourage young people to major in mathematics.