Human-Compatible AI for Well-being

Three Bachelor students from the new FHNW School of Computer Science gave a presentation at the AAAI 2025 Spring Symposia. Şahan Hatemo, Christof Weickhardt, and Luca Gisler were on site in San Francisco on April 1, 2025 and presented the paper “Revisiting the Trolley Problem for AI: Stereotypes and Biases in Large Language Models and their Impact on Ethical Decision-Making” at the symposium “Human-Compatible AI for Well-being: Harnessing Potential of GenAI for AI-Powered Science”. Prof. Dr. Oliver Bendel from the FHNW School of Business, the fourth author, gave them feedback on the paper and located the project in the discipline of machine ethics. Şahan Hatemo also presented the paper “Miss Tammy as a Use Case for Moral Prompt Engineering” by Myriam Rellstab and Oliver Bendel, who were unable to be in San Francisco. For the technology philosopher, the chatbot is an implementation of machine ethics methods in the field of Large Language Models (LLMs). He speaks of moral prompt engineering, by which he means the combination of prompt engineering and, where appropriate, retrieval augmented generation (RAG). The proceedings of the AAAI 2025 Spring Symposia will be published shortly.

Fig.: The participants of the symposium