The academic peer review system desperately needs a redesign. We’re approachin… · @noahatwhere · TwKit
@noahatwhere
The academic peer review system desperately needs a redesign. We’re approaching a point where AI writes the paper, AI reviews the paper, and AI helps authors revise it to satisfy another AI. And here’s the absurd part: if reviewers increasingly rely on LLMs, researchers may be incentivized to write papers that score well with LLMs rather than papers that contribute meaningful science. You could spend years doing rigorous research, only to be outscored by someone who knows how to optimize their manuscript for an AI reviewer. We’ve essentially created a system where researchers are learning to game the judgment of machines that were supposed to help us evaluate human knowledge. At some point, we’re no longer optimizing for scientific discovery. We’re optimizing for AI approval. Peer-reviewed, by nobody.
曝光 29 · 评论 0 · 点赞 0 · 书签 0
查看原帖