Conceptual Design and Implementation of Training Scenarios for Human-Centered Artificial Intelligence in Learning Factories

The number of AI solutions within industry and especially manufacturing has been steadily increasing in recent years. Consequently, AI diffuses into many different areas of employees’ everyday work life. In order to avoid a mismatch or conflict within employee-AI-interactions, a human-centric development of such technology is required. For instance, AI systems need to be adaptable to specific work situations. At the same time, employees need to be able to operate and manage these adaptable AI systems, which requires corresponding expertise. Establishing according skills represent the decisive factor to keep up not only with current but also future technology. However, approaches dedicated to learning how to handle AI have yet to be developed. We fill this gap by establishing a training concept within a learning factory for human-centered adaptive AI-work systems using adaptable demonstrators. The concept is part of an extensive training series and is based on a didactic model. Within this scenario, we follow a socio-technical perspective by considering human, technological, and organizational changes at different levels of observation that result from the introduction of AI in the work system. The concept is implemented within the learning and research factory of the Chair of Production Systems at Ruhr-University Bochum. In this paper, we contribute not only to science but also to practice by deriving action-oriented implications and the description of numerous demonstrators. The development and validation of the learning scenario takes place within the BMBF-funded project human-centered AI network (HUMAINE)