The Future of Work in Germany (FoWiG)
Research leader:Fregin, M
This research project empirically assesses the consequences of implementing artificial intelligence for employment, productivity, quality of working life, and skills requirements of jobs. We study how ai transforms the work of human experts in various roles and functions in large German companies.
The study is based on empirical evidence gathered using a mix of qualitative and quantitative research methods, including expert interviews for selected occupations and fields of activity, employer surveys, field experiments that model the causal effect of ai implementation on outcomes, and quantitative analyses of administrative data. aim is to conduct a representative survey for IBM employees.
At Siemens, we studied the implementation of an IBM Watson-based ai that serves as a chatbot (named: Carl) to answer questions related to human resources. We performed quantitative analyses of administrative data, as well as qualitative analyses of in-depth interviews. We found no substitution effects in our quantitative analyses, but qualitative analyses of interviews with workers in managerial roles suggest that such effects may still occur later on in the process.
At Deutsche Telekom Service GmbH, we studied the effects of the implementation of a smart technology (PIA) that aids workers in a customer services department to answer the complex questions that customers have. We applied a randomized control field experiment that enables us to identify the causal effects of PIA. There are two main objectives of the use of PIA: to increase the quality of order processing and service quality, and to increase in employee satisfaction. Preliminary analyses of the RCT suggest that there are no positive effects of PIA on employee satisfaction, and also not on the quality of order processing.
The FoWiG research project has been developed in collaboration with IBM, ver.di, and Input Consulting, with support of the Federal Ministry of Work and Social Affairs. Next to conducting the research, our scientists have been engaged in active dialogue sessions with representatives of the German federal government, the AI observatory, large companies, and unions, to help co-create policies that are aimed at maximising returns to artificial intelligence, while minimizing potential negative outcomes for workers.