This study explores the relation between risk of job automation and participation in adult education and training (AET) and examines variation in that relation across welfare regimes distinguishing between situational and institutional barriers. Using microdata of PIAAC, we analyze participation in formal or nonformal AET for job-related reasons in relation to the risk of automation of the respondents' occupation after controlling for main sociodemographic characteristics. Logistic regression models are run on respondents from 14 European countries representing different welfare regimes: Denmark, Norway, and Sweden (Scandinavian countries); Italy, Greece, and Spain (Southern European); Czech Republic, Slovakia, and Poland (Central and Eastern Europe), Belgium, France, and Germany (Continental); and United Kingdom and Ireland (Anglo-Saxon countries). Our findings confirm that workers in occupations at high risk of automation were found to be consistently less likely to participate in job-related AET, quite irrespective of welfare regime. [The paper was presented at XIII Conferenza Espanet Italia--Il welfare state di fronte alle sfide globali (Venezia, 17 September 2020).]