Reasoning and Planning with Dynamically Changing Norms

In three linesPaper proposing an approach to guide AI agent planning with dynamically changing norms in human-AI settings. Contributions: a defeasible calculus for resolving normative conflicts and using such norms as guardrails on plans. Theoretical validation via formal proofs and empirical validation with SocialBot on natural language dialogue tasks.

Summary generated by Claude — human-verified