Third‑Party Disclosure of AI Agent Participation

Determine whether, in interactions involving AI Agents—autonomous software systems that act on a user's behalf—third parties must be informed that an AI Agent is acting and whether such disclosure should be legally mandated under agency law to enable appropriate evaluation of responsibility and performance.

Background

The paper contrasts computer science and legal conceptions of agency, noting that computer science’s model of agents does not explicitly address third-party knowledge, whereas legal agency emphasizes disclosure so third parties can assess responsibility and performance. While APIs, authentication, and existing ecommerce practices mitigate many third‑party risks, the authors identify a residual uncertainty about disclosure requirements when AI Agents are involved.

Within their broader argument about combining socio‑legal norms with technical mechanisms to build responsible AI Agents, the authors raise a direct question about whether third parties need to be told when an AI Agent—not a human—is acting, highlighting a gap that current technical safeguards do not fully resolve.