Scientific writing in the age of AI

As researchers across disciplines continue to explore the potential benefits and applications of artificial intelligence (AI), Yale physicist Chiara M.F. Mingarelli recently tested the use of AI in scientific writing—and found both strengths and weaknesses.

While attending a two-week conference on the discovery of the gravitational wave background and its large amplitude, Mingarelli, an assistant professor of physics in the Faculty of Arts and Sciences (FAS), decided to use AI to create a summary of the conference. She uploaded conference transcripts to an AI platform and reviewed each new summary iteration with colleagues.

Mingarelli said that with each new iteration, the summary grew more detailed. But its deficiencies were "glaring," she noted. At times it misinterpreted complex discussions, while also lacking in coherence and depth.

"At best, the document resembled a poorly written conference proceeding," Mingarelli wrote in an op-ed describing the experience for the journal Nature Astronomy. "At worst, a tabloid version of our meeting."

More information:
C. M. F. Mingarelli, Scientific writing in the age of AI, Nature Astronomy (2025). DOI: 10.1038/s41550-025-02493-y

Provided by Yale University