International
A new analysis of over 41 million scientific papers and many of the authors' careers suggests using AI tools in research is linked to career success. Researchers used a language model to identify AI-augmented research published from 1980 to 2025 in natural sciences subjects, where scientists may use tools like machine learning and generative AI, but don't typically research AI methods directly. Authors of "AI-augmented" research published 3 times as much, got nearly 5 times the citations, and progressed faster in their careers. However, papers that didn't use AI covered a wider range of topics and tended to cite each other more. The authors say the use of AI could be encouraging researchers to "converge on the same solutions to known problems" rather than coming up with new ones, narrowing scientific exploration and engagement.
Note: Not all attachments are visible to the general public. Research URLs will go live after the embargo ends.
Sign In
