Research
AI-written peer reviews outscored top human experts in blind comparisons, Harvard working paper finds
PsyPost · By Eric W. Dolan ·
The Brief
- A Harvard working paper reports that advanced AI models produced peer reviews of unpublished academic papers that scored higher than those of well-paid human experts in blind comparisons.
- The blind comparisons were judged by an AI model and by human evaluators, according to the PsyPost summary.
- The study covers reviews generated under ideal conditions for the human experts, per the source; the paper itself is a working paper and not yet peer reviewed.
Why it matters
If AI-written reviews are judged more thorough and constructive than expert reviews, journals and conferences may need to rethink how review quality is assessed and who writes reviews.
Who’s affected
- researchers who submit or review manuscripts
- journal editors and conference organizers
- academic publishers
A new study suggests that advanced artificial intelligence models can generate more thorough, accurate, and constructive peer reviews of unpublished academic papers than well-paid human experts operating under ideal conditions. In blind comparisons judged by an AI model and by…
Timeline
PsyPost reports that a Harvard working paper finds AI-written peer reviews outscored top human experts in blind comparisons
PsyPost – Psychology News
AI-written peer reviews outscored top human experts in blind comparisons, Harvard working paper finds