AI-written peer reviews outscored top human experts in blind comparisons, Harvard working paper finds
Artificial intelligence may now write more useful feedback on unpublished research papers than well-paid human experts. In blind comparisons, both an AI judge and former journal editors preferred the AI-written reviews in the large majority of cases.
A new study conducted by Harvard University researchers suggests that advanced artificial intelligence (AI) models can produce peer reviews for unpublished academic papers that are more comprehensive, accurate, and constructive compared to those generated by highly skilled human experts. In a blind comparison test, where both AI and human reviewers evaluated the same manuscripts, the AI-written reviews emerged victorious in the vast majority of instances and exhibited a significantly lower rate of incorrect claims, according to a working paper published by Harvard researchers.
This finding raises concerns about the future of scientific publishing and the potential impact of widespread reliance on AI on the quality of scientific research.
Written by urgent.news from PsyPost's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
