Take-away: We study how knowledge-driven RAG can improve trustworthiness of agricultural AI advice. Part of the EU-FarmBook project.
Corresponding Paper (WIP)
Take-away: We study how knowledge-driven RAG can improve trustworthiness of legal AI advice. Part of the Norm Engineering project.
Corresponding Paper (WIP)
Take-away: We introduce an uncertainty-driven RAG (CERTA) and show hwo it answers more approprately on in QA tasks, answering "I don't know" more often in the right moments.
Corresponding Paper
Take-away: We propose that the gap between functional and normative Trustworthy AI should be bridged with a descriptive framework.
Corresponding Paper
Take-away: We introduce PACCART, a framework for privacy protection through multi-agent argumentation.
Corresponding Paper