PNAS Study Finds Generative AI Use in Grant Applications Reduces Topic Uniqueness

Written by

in

A PNAS study found that researchers who used generative AI to write federal research grant applications proposed topics more similar to recently funded projects. An analysis of data from 2021 to 2025 showed that when AI involvement increased from the 25th percentile to the 75th percentile, topic uniqueness declined by approximately 5 percentage points for NSF-funded projects and 4 percentage points for NIH-funded projects. Among NIH applications, those with greater AI involvement were approximately 4 percentage points more likely to receive funding and subsequently produced about 5% more papers, although there was no significant difference in the number of highly cited papers. The study was observational and cannot establish causation.

Source:PsyPost · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *