AI-written papers authors don't understand are flooding arXiv, warns Dietterich
begusgasper · x · 2026-09-14
Thomas Dietterich flags a new trend in arXiv (and presumably conference/journal) submissions: papers whose contents the authors likely do not understand. Quoting him, Nina Begus muses that we may be the last generation to write the human science canon — future generations will view our output the way Egyptians regarded their pyramids, wondering how it was done and lamenting the lost know-how.
The thread's core concern: AI-assisted writing is decoupling paper production from author understanding, threatening the peer-review and quality-control mechanisms of science.
More from AGI Musings
- Nat Lambert: Open models could be what holds the AI ecosystem together — natolambert · 2026-10-31
- Kai-Fu Lee's 'AI Native' ships Nov 3 with endorsement from Satya Nadella — kaifulee · 2026-10-31
- AI security's dark age: agents can probe every endpoint and attack surface — graceisford · 2026-10-31
- AI dinner takeaways: orgs shipping 5x faster, yet GDP isn't growing 30% a year — graceisford · 2026-10-31
- Companies that replaced engineers with AI are rehiring seniors at a premium — DavidLinthicum · 2026-10-31
- Nathan Young proposes prediction markets and AI reporting for AI-community accountability — NathanpmYoung · 2026-10-31