UNU-MERIT Working Papers Series
Does Training in AI Affect PhD Students’ Careers? Evidence from France
Pierre Boutros, Eliana Diodati, Michele Pezzoni & Fabiana Visentin #2025-016The rise of Artificial Intelligence (AI) urges us to better understand its impact on the labor market. This paper is the first to analyze the supply of individuals with AI training facing the labor market. We estimate the relationship between AI training and individuals’ careers for 35,492 French PhD students in STEM who graduated between 2010 and 2018. To assess the unbiased effect of AI training, we compare the careers of PhD students trained in AI with those of a control sample of similar students with no AI training. We find that AI training is not associated with a higher probability of pursuing a research career after graduation. However, among students who have AI training during the PhD and pursue a research career after graduation, we observe a path dependence in continuing to publish on AI topics and a higher impact of their research. We also observe disciplinary heterogeneity. In Computer Science, AI-trained students are less likely to end up in private research organizations after graduation compared to their non-AI counterparts, while in disciplines other than Computer Science, AI training stimulates patenting activity and mobility abroad after graduation.
JEL Classification: J24, O30
Keywords: Artificial Intelligence, Training, PhD students’careers
DOI: https://doi.org/10.53330/EPFW4463
Download the working paper (PDF)
