# AI Usage in PhD Dissertations Linked to Industry Career Paths STEM doctoral graduates using AI for writing are increasingly pivoting toward industry over academia. - **Source:** [Nature](https://brifing.si/source/4d850c3b-dd79-4024-bc91-db2571889199) - **Published:** 2026-10-16 - **Topic:** [People & Culture](https://brifing.si/c/people) - **Original article:** - **On Brifing:** https://brifing.si/s/0cf346c2-ff08-4adc-ad05-523eac6fd06d ## Rising AI Adoption Trends A working paper analyzing over 226,000 STEM doctoral dissertations reveals a sharp increase in AI-generated content. While such text was absent before 2023, it surged to 29.4% of abstracts by early 2026. Mathematical and computer sciences lead this trend, with 36% of abstracts showing evidence of AI assistance during the research process. ## Academic Versus Industry Trajectories Data suggests a correlation between AI usage in dissertations and post-graduation employment choices. Graduates who utilized AI tools were 17% less likely to pursue academic roles compared to their peers. Conversely, these individuals were 16% more likely to secure positions within the industry, signaling a potential shift in long-term career professional focus. ## Geographic and Linguistic Variations The study highlights distinct patterns based on the educational background of the graduates. Approximately 41% of dissertations from students educated in non-English-speaking countries contained AI-generated abstracts. This contrasts with 23% for those from English-speaking nations. Furthermore, the link between AI usage and reduced academic employment was less pronounced for international undergraduate students. ## Methodology and Detection Limits Researchers utilized the Pangram service to evaluate abstracts submitted between January 2019 and May 2026. The study categorized abstracts as either AI-written or not, without distinguishing between partial or full generation. However, the authors noted that 90% of the flagged abstracts were classified by the detection tool as being entirely AI-generated. ## Concerns Over Research Depth Economist Daniel Gross questions whether students delegating writing tasks to AI maintain the necessary depth of understanding required for scientific rigor. The concern is that if students perform only half the work, their ability to function as high-level scientists in uncertain or incomplete research areas may be compromised during their professional careers. ## The Challenge of Uncertain Knowledge Engineer Ang Liu emphasizes that generative AI is highly effective for synthesizing existing knowledge but struggles with unexplored research frontiers. Rigorous training becomes essential when research moves into areas where answers remain unknown. Relying on AI for writing may inadvertently bypass the critical thinking processes required to navigate these complex, uncertain scientific landscapes. ## NBER Working Paper Insights The National Bureau of Economic Research study aims to understand how AI integration shapes future scientific capabilities. By analyzing profiles on LinkedIn, researchers mapped the career outcomes of graduates against their dissertation writing habits. This analysis provides a window into how technological tools are fundamentally altering the training and professional trajectory of scientists. ## Future of Scientific Training The shift toward industry roles among AI-using graduates suggests that academic institutions may need to re-evaluate their training models. As AI becomes a standard tool in research, the definition of scientific proficiency is evolving. Understanding these patterns is crucial for shaping the next several decades of scientific inquiry and the professional development of researchers. ## Sources covering this story - https://www.nature.com/articles/d41586-026-03253-2 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._