This study, carried out by CEDEFOP (European Centre for the Development of Vocational Training), uses a new dataset based on online job ads – OVATE for Cedefop skills – with information on skills and work activities requested by EU employers.
The data provides an overview of the task profiles of detailed professions facing a higher risk of automation or professions using alternative digital technologies (robots, computer software, artificial intelligence). The document explores appropriate models of machine and deep learning to test how well a decent set of task indicators can anticipate the automation of the profession. The professional activities associated with a higher risk of automation at work and exposure to robots (e.g. equipment control, physical activity), usually concentrated on routine or manual work, differ from activities that are significant in occupations with higher exposure to AI (e.g. creative thinking, standard assessment).