Who uses which AI? Educational inequalities in workplace ai use among workers at different career stages

Abstract

Artificial intelligence (AI) may reshape workplace inequality, yet little is known about how education shapes actual AI use. This study compares employees with vocational training and university degrees, arguing that occupation-specific skills and transferable analytical and adaptive abilities may lead to different patterns of AI adoption. We distinguish AI requiring general competencies, such as chatbots and image generators, from systems embedded in occupational processes, such as guidance systems and collaborative robots. Using National Educational Panel Study (NEPS) data on 1,867 employed individuals, we examine educational differences in AI use, variation across career stages, and the role of pre-existing task profiles. Linear probability models show that university graduates more frequently use AI requiring general competencies, particularly later in their careers.
Vocationally trained employees do not use occupation-specific AI more frequently. Pre-existing task profiles partly explain educational gaps, while AI use accounts for a small share of the wage gap. Overall, workplace AI use reflects established inequalities in the German labor market.

Chair 

Paolo Barbieri, Università di Trento

Discussant

Beatrice Caniglia, Università di Trento

Registration 

Send an e-mail to csis@unitn.it by 18 November 2026, 23:59.