In this project we develop new measures for the advancement of AI and robotics in Europe using Natural Language Processing techniques on patent data from the European Patent Office. Combining the patent-based measures with administrative data on establishments in Germany, we find negative employment effects for AI whereas robots negatively affect low-skill employment but have positive effects on medium- and high-skill employment. Wage effects of AI are positive for all skill groups while robot exposure is associated with lower wages. 

We also investigate the employment and wage responses in local labor markets using a shift-share design. For robots, we find negative employment effects, especially in manufacturing. AI has negative employment effects in both manufacturing and services. Robots replace mostly low-skilled workers both inside and outside of manufacturing and AI has negative employment effects for low- and medium-skilled workers in manufacturing and services.

How does artificial intelligence (AI) affect the task content of jobs, and how do exposed workers adjust? We combine a novel patent-based measure of AI exposure with survey data on tasks and administrative data on worker careers to provide new answers to these questions for Germany. Unlike the previous wave of robots, AI exposure shifts the content of jobs from non-routine abstract tasks towards routine tasks. This task shift occurs mainly within narrowly defined occupations within manufacturing and  services and is stronger among low- and medium-skilled workers. AI exposure also leads to substantial reallocation of workers from exposed to non-exposed industries. 

The debate on the impact of digital technologies on the labor market has so far primarily focused on aggregate employment effects, while the impact of digitalisation on the quality of jobs is less studied. In this project I examine the individual-level relationship between exposure to digital technologies and job quality. Additionally, I explore the mediating role of employer-provided training and personnel management. To do so, I use a linked employer-employee survey and administrative data from Germany. In fixed-effects regressions I find that exposure to advanced digital technologies, such as AI, is associated with improved working conditions and increased participation in employer-provided training. Conversely, high exposure to basic digital technologies, like computers and computer-controlled machines, is associated with lower job quality and training participation. These effects are more pronounced among male and older employees, underscoring the varied impacts of digital technology exposure across demographic groups.