AI-written PhD abstracts linked to industry exit from academia

Roughly 30% of US STEM PhD dissertations filed in the first five months of 2026 contained AI-generated text in their abstracts, Nature reports, citing a working paper by Daniel P. Gross, Dror Shvadron and Hansen Zhang. 'AI writing is absent before 2023, present in 29% of dissertations filed in 2026, and rapidly growing,' the authors write.

The team scored abstracts from more than 226,000 STEM dissertations completed between January 2019 and May 2026 at US universities, using the Pangram detection service and cross-checking the signal against stratified samples of full texts.

Adoption is uneven. 'Math and computer sciences led AI use at 36%; dissertations by non-English-speaking country graduates were flagged 41% of the time versus 23%,' with higher rates also at lower-ranked institutions.

The sharper finding is the career divide. Among same-program, same-cohort peers, graduates whose abstracts tripped the detector were less likely to enter academic research and more often landed in non-tenure-track roles or industry, with the pattern strongest among US-origin students. For that group, AI-flagged abstracts tracked with being '17% less likely to pursue academic employment and 16% more likely to work in industry' than peers without flagged use. Two of the researchers we track in our Who's Who directory shared the Nature write-up inside a day.

For what it is worth, the acknowledgments note that 'the researchers themselves employed Claude for initial drafting of discussion sections and portions of the data appendix,' with the AI-generated passages substantially revised by hand.