Study of 227K STEM PhD abstracts links AI use to leaving academia for industry
ValerioCapraro · x · 2026-10-06
A new preprint analyzes nearly 227,000 American STEM PhD dissertation abstracts (2019–May 2026), using Pangram to detect AI usage.
Key findings:
- AI use correlates with a lower likelihood of staying in academia and a higher likelihood of entering industry, especially among US-origin graduates
- Those who stay in academia are more likely to hold non-tenure-track roles
The authors caution this is correlational: students already planning to leave may simply be more inclined to use AI. But Valerio Capraro highlights a deeper concern — writing forces you to test whether your argument works, and offloading it may erode the very skills needed to become a good scientist.
Related event: Study of 227K PhD Theses Links AI Use to Leaving Academia(3 posts)→
More from AGI Musings
- dbasch: A dog feels, Claude doesn't — Claude is a brand, not a being — dbasch · 2026-10-06
- Proving Famous Conjectures Will Become the Cheapest Science, Math Paper Organizer Argues — littmath · 2026-10-06
- AI Math Breakthroughs Prompt Question: How Many Real Engineering Problems Have Math Solutions? — Afinetheorem · 2026-10-06
- Agents' Ability to Act Makes Them Useful and Dangerous: Safety Is Now an Ops Problem — alifcoder · 2026-10-06
- AI detector triggered a French literary scandal: probability became conclusion — moniquejmorrow · 2026-10-06
- Researcher laments how slow humans look as algorithm design breakthroughs pile up — michaelchchoi · 2026-10-06