US: the proposed Stop Rogue AI Act aims to give companies and agencies back control over AI agents

The Stop Rogue AI Act (H.R. 10362) was introduced in the US House of Representatives on 14 September 2026. It was introduced jointly by Democrat Josh Gottheimer and Republican Mike Lawler. The bill responds to the fact that organisations often do not know how many AI agents are running in their systems, who launched them and what they have access to. Agents may be launched by employees, embedded in third-party software or deployed by vendors.

The National Institute of Standards and Technology (NIST) is to issue, within one year, standards, guidelines and best practices for discovering, verifying and controlling AI agents. Under them, organisations should maintain a continuous, machine-readable inventory of all agents. They should verify agents' identity through cryptographically verifiable mechanisms and be able to allow, deny or revoke an agent's access at any time.

The standards are to be binding on federal agencies and their contractors, through an amendment to the federal procurement rules. For the rest of the private sector they will remain voluntary.

The bill has been referred to the Committee on Science, Space, and Technology and the Committee on Oversight and Government Reform.