To determine whether to follow a particular command, a RaD-AI agent must be able to assess what is the core objective of the human(s) it serves and align its values accordingly.  For this purpose, knowing the human's current plan is highly relevant, thus making methods from the plan recognition and goal reasoning communities potentially useful. We welcome submissions directly addressing RaD-AI, as well as other topics relevant for the development of RaD-AI, including but not limited to: 

Intelligent Social Agents

• Goal Reasoning

• Plan Recognition

• Value Alignment

• Social Dilemmas 

• Norm Interpretation and Learning

Human-Agent Interaction

• Human-agent Trust

• Interruptions

• Deception

• Command Rejection

• Explainability 

Societal Impacts

• Legal and Ethical Reasoning

• Liability

• AI Safety

• AI Governance