Engaging in Dialogue about an Agent's Norms and Behaviors
This addresses the problem of transparency and control in AI systems for users interacting with agents governed by complex norms, though it appears incremental as it builds on existing planning and temporal logic frameworks.
The paper tackles the problem of enabling agents with moral and social norms to communicate about their behaviors in natural language, allowing users to query, modify, and test norm effects. The result is a set of capabilities that facilitate interactive dialogue between humans and norm-governed agents.
We present a set of capabilities allowing an agent planning with moral and social norms represented in temporal logic to respond to queries about its norms and behaviors in natural language, and for the human user to add and remove norms directly in natural language. The user may also pose hypothetical modifications to the agent's norms and inquire about their effects.