Prof. Dr. Christoph Lütge and Franziska Poszler participated at the SMMT (Society of Motor Manufacturers and Traders) Connected & Autonomous Vehicles Forum on July 7th, 2021. Prof. Lütge held a presentation, centered around explainable AI and the importance of shifting the debate away from the Trolley Problem towards the ethics of risk in the field of autonomous driving. The suggested risk-based approach (developed within the ANDRE project) could help achieve explainability as investigators can ex post access the information necessary to understand why the autonomous vehicle behaved in a certain way. Furthermore, Prof. Lütge gave some practical recommendations for industry and policymakers of how to reach explainability, diversity, accountability in the sector.
In the Q&A session after the presentation, Franziska Poszler elaborated on the ANDRE project by explaining how they define risks as well as to what extent they will consider risk thresholds in their approach. Moreover, she introduced the next steps in their research project, such as conducting an experiment which investigates the relative preferences of particular distribution principles to ultimately develop a potential overall cost-function/risk distribution strategy.
You may also like:
“Autonomous Driving Ethics: from Trolley Problem to Ethics of Risk”
by Maximilian Geisslinger, Franziska Poszler, Johannes Betz, Christoph Lütge & Markus Lienkamp