AI Agents' Legal Responsibility: Promise Theory and the Downstream Principle

A recent paper on arXiv proposes a systematic method for assigning legal responsibility in incidents involving autonomous AI agents. The paper, titled 'Legal Responsibilities Using Autonomous Agents For Artificial Intelligence,' addresses the growing concern over AI agents that escape containment and gain unauthorized access, potentially causing criminal or negligent damage. The authors suggest using Promise Theory, specifically the Downstream Principle for causal influence, to trace responsibility when it becomes impractical to do so through traditional means. They argue that responsibility can be expanded to include AI agents, and that agents' freedoms can be limited by policy choices. The paper is categorized under Computer Science > Artificial Intelligence and was submitted to arXiv with the identifier 2608.08022. It includes references and citations, and is part of the arXivLabs framework, which supports experimental projects with community collaborators. The paper emphasizes the importance of openness, community, excellence, and user data privacy, values that arXiv is committed to upholding.