Overview
- Prime Minister Kristen Michal announced the proposal on Wednesday to issue unique state-backed ID numbers for AI agents so they can be identified and authorised separately from their human principals.
- The stated aim is to give agents narrow, controllable rights such as read-only access, document drafting only, or capped financial authority to prevent broad use of a person’s credentials.
- Reporters say AI agents today often operate using their owners’ logins for email, cloud storage and banking, which lets the software act with the same rights as the person and creates risks of data loss, unauthorised transfers and data exfiltration.
- The measure is a government proposal, not yet law, and key details remain unresolved including whether IDs would be mandatory, who would bear liability for agent mistakes, the technical design for granular controls and any rollout timetable.
- Estonia’s near-universal e‑ID use and e‑Residency programme make the issue urgent there and give the country a practical testing ground that could shape how other states regulate AI agents and digital access.