In the law of artificial intelligence, the legal handling of innovations is the focus of research interest: self-learning systems end the monopoly of humans on rational decisions made under conditions of uncertainty. Technical products and systems that adapt to changing environmental conditions without human intervention will lead to a redistribution of liability burdens between manufacturer and user, possibly also with the addition of new bearers of liability law responsibility. In addition, the question arises under which conditions a self-learning system is to be qualified as defective at all.
This research focus was worked on within the framework of a sub-project led by Janine Wendt from 2016 to 2021 together with colleagues from mechanical engineering and mathematics in the Collaborative Research Center (SFB) 805.
The results from the SFB flow into the pilot project RoboTrust of the Zentrum für verantwortungsbewusste Digitalisierung (ZEVEDI) (Centre for Responsible Digitalisation), which is dedicated to the trustworthy design of human-robot interaction – also in great interdisciplinarity with colleagues from computer science, among others. Anthropomorphic service robots have already been used for years. There is potential in customer service as well as in professions where human service providers are overworked due to the shortage of skilled workers. Through their special physical presence, anthropomorphic robots can establish an emotional bond with human users. However, the connection between software and machine also gives robotics the power to physically violate bodies and property.
Janine Wendt's sub-project explores the foundations for authentic and acceptable human-robot interaction, taking into account the aspects of responsibility and liability, while Prof. Indra Spiecker's sub-project is dedicated to privacy.
The findings from both research projects will be processed in two comprehensive publications, on the one hand in the major commentary on the EU regulation establishing harmonised rules for artificial intelligence (so-called Artificial Intelligence Act), editorship (together with Prof. Dr. D. Wendt), Nomos Verlag, 1st edition 2024, as well as in a supplementary handbook on the law of artificial intelligence in the European Union (also together with Prof. Dr. D. Wendt), Nomos Verlag, 1st ed. 2024. The publications will be available in German and English.
