Researchers warn AI-generated paper critiques are the wrong way to do academic criticism
TaniaBabina · x · 2026-10-29
- An AI-generated analysis claiming errors in economists Emmanuel Saez and Gabriel Zucman's research circulated via @MargRev.
- @sccath pushes back: AI isn't good at this kind of assessment, and the prompt used is unknown.
- Some "errors" are just criticisms from papers published after the analyzed ones — a normal part of science.
- Takeaway: this is not the right way for researchers to criticize each other.
More from AGI Musings
- Why AI-generated technical slides are so bad: novices mistake slides for the talk — tallinzen · 2026-10-29
- Reuters Institute's Felix Simon: conversational AI has arrived as an information intermediary — _FelixSimon_ · 2026-10-29
- Reuters Institute 9-country survey: weekly AI use jumps from 34% to 51% in a year — _FelixSimon_ · 2026-10-29
- Scoring 342 US jobs for AI risk: office work at 94%, care jobs at 17%, and riskier jobs pay more — albertsimondev · 2026-10-29
- ChatGPT May Make This Humanity's Most Important Decade, as Capital Fuels Gargantuan Training Runs — abhiadesai · 2026-10-29
- Google Research VP: By 2030 Every Researcher Could Have Their Own AI Lab — dumpshoot · 2026-10-29