Legal Authority for Autonomous AI Agents

The first question any agent must answer is simply “Who are you?” But this is not a trivial inquiry. For an AI‑based agent, establishing identity requires more than a label or a model ID. It requires a resolvable, unique identifier that allows relying parties to cryptographically verify which specific agent instance they are interacting with; it also requires a clear statement of the ultimate beneficial owner—the human user or legal person in whose name the agent acts and who carries legal accountability for its behaviour—together with the operator or provider responsible for safety, security and correct functioning. eIDAS 2.0 supplies the legal and technical primitives for this binding through the concepts of personal identification data, electronic attestations of attributes, and their qualified forms, allowing issuers and relying parties to attach legal effect and assurance levels to digital identities and roles in a harmonised way across the Union.