What would it actually take for an AI system to be reason-responsive, to act, prefer, and decide for reasons in a way we’d recognise as normatively meaningful? That’s the question at the heart of NoRA, our one-day workshop bringing together researchers from philosophy and computer science to explore the intersection of normative reasoning and machine learning.

On the agenda:
·      the structure of practical and epistemic reasons
·      how neural networks might be explained in terms of reasons
·      the coherence of AI preference orderings
·      what genuine reason-responsiveness in AI systems would require

Philosophers and ML researchers working the same problems from different directions, in the same room. Exactly the kind of exchange responsible AI needs more of.

Organised by André Steingrüber, Yannic Muskalla and Kevin Baum from the Responsible AI and Machine Ethics (RAIME) group at Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI), as part of CERTAIN.

More about NoRA: https://lnkd.in/eQw4bb_Z