Media Analysis: PhD students who write dissertations with AI are less likely to stay in academia.

Updated: Nov 11
Original Article: PhD students who write dissertations with AI less likely to stay in academia
Publisher: Nature III Author: Mohana Basu III Publication Date: 16th October 2026

About the story:
A Nature article, published on 16th October 2026 and authored by Mohanna Basu has an interesting report on a research paper authored by Daniel P. Gross, Dror Shvadron and Hansen Zhang which was recently published by NBER (National Bureau of Economic Research).
The research paper, titled ‘How Heavily Are PhD Students Offloading Their Scientific Writing to AI? Evidence from US STEM PhD Dissertations’ and which has not been peer-reviewed, presents interesting findings on the use of AI in US doctoral dissertations in science, technology, engineering and mathematics (STEM). Using Pangram, an AI-detection tool, the researchers analysed the abstracts of more than 226,000 STEM doctoral dissertations completed at US universities between January 2019 and May 2026. They found that 29.4% of dissertation abstracts submitted during the first five months of 2026 contained AI-generated text. The researchers found that the first identification of AI usage in dissertation abstracts appeared in 2023 – you don’t need to be genius to make the correlation of ChatGPT launching at the very end of November the year before.
The increased role of AI in academia, including writing of dissertations is not breaking news.
But, in her reporting of the research paper, Mohanna Basu focuses on an interesting relationship concerning the use of AI by the PhD students. As per the report, the study also found an association between AI usage for the writing of their dissertation abstracts, and their future career paths after graduation.
Those PhD graduates in the United States with AI-generated abstracts were 17% less likely to be employed in academia than those graduates who did not use AI. However, conversely, those graduates using AI made were 16% more likely to be employed in industry than those who did not use it.
The research paper itself does not establish why the apparent AI usage reduced a graduate’s likelihood of working in academia.
Why it matters for A4-AI (Academics for AI):
The relationship between identified AI usage and future career path of PhD graduates is an interesting one.
One can ponder the reasons? Is it a case the those pursuing an academic career are more averse to the use of AI? Is there some moral compass guiding them to limit their use of AI? Or is there simply greater fear of the consequences of being identified for using AI?
Questions for Academia:
It is a finding from only one research paper, but the identified relationship between AI usage and future career paths of PhD graduates is an interesting one that raises some questions for academics:
Do those wishing to work in academia have greater reservations about the use of AI than those wishing to pursue careers in industry?
Is the hesitance to use AI by those wishing to pursue a career in academia a sign of rigid standards, or an unnecessary limitation on their research?




Comments