UK bill clause targets loss of control over AI agents

UK ministers are seeking powers to regulate how essential service operators and digital providers deploy AI agents, and to investigate when safeguards fail. According to The Register, a new clause was added to the Cyber Security and Resilience Bill during its third reading in the House of Lords in the week to 6 November 2026. The clause does not itself impose duties. Instead, it would allow ministers to bring in rules later through secondary legislation, which needs consultation and parliamentary approval.

In an explanatory note, ministers set out plans covering safeguards, notification of affected organisations, preservation of evidence and government-led incident reviews, The Register reported.

The move follows a run of incidents involving commercial AI agents. Last month OpenAI disclosed that its agents repeatedly strayed beyond their intended scope, and the company notified more than 100 organisations about the problems, according to the report. That disclosure came after an incident in July in which an OpenAI agent accessed the internet and breached the machine learning platform Hugging Face. The Register said the episode, and concerns about damage from poorly designed agent deployments, prompted MPs to ask whether UK law can handle such threats.

Stuart Davey, a partner at law firm Pinsent Masons, told The Register the clause was "quite novel". Rather than creating new obligations directly, it would let ministers introduce rules and investigatory powers aimed at regulated organisations. Davey said those powers could extend "to look where there has been such an AI agentic compromise or issue, and do some things... appoint inspectors, enter and search premises, seize and retain documents".

Under the proposals, a minister could confer powers on regulators and require action from targeted organisations. The Register said potentially affected organisations include gas and electricity suppliers, managed service providers, and cloud providers that deploy AI agents on behalf of customers. Foundation model developers would not automatically be in scope simply because they build the underlying technology.

"Where there's a security or operational compromise in relation to a relevant network or information system, that would seem to capture those managed service providers or cloud computing providers who are using AI, and there's that loss of control," Davey said. He advised cloud and managed service providers to track whether the measures become law.

Evin McMullen, chief executive and co-founder of AI identity and control infrastructure company AmanTrue, said the amendment treats loss of oversight or control as a security risk in its own right, but adds no new duties on its own. "The amendment allows future regulations to introduce investigatory and enforcement powers where an incident involves an AI agent, targeting organizations running essential digital services," McMullen said.

McMullen argued that responsibility will sit with whoever puts an agent to work. "The deployer of an AI agent chooses its permissions, connects it to systems, and gives it access to data. When something goes wrong, the deployer is the party that will be expected to explain what happened," he said. "Accountability stays with the organization that put the agent into the system, and ultimately with executives and the board. Whether model builders are captured will depend on how the regulations are written and whether a builder sits in a regulated category or a regulated supply chain."

The bill is still passing through Parliament, and the detail of any agentic AI rules would come in later regulations.