Talk at Cambridge Political Psychology Lab: Moral Alignment for Agentic AI Systems
Date:
Gave a 1-hr talk dsicussing Moral Alignment for Agentic AI Systems. In this talk I discussed agency in AI, existing approaches to Alignment (as described in our preprint), and my work on training or fine-tuning RL and LLM agents with intrinsic rewards, with particular focus on the unexpected findings from our papers at IJCAI’23, AIES’24 and the latest preprint.