Britain's minister for artificial intelligence, Kanishka Narayan, warned frontier AI laboratories on Monday that they should not release systems whose risks they cannot adequately control. Speaking at Imperial College London, he also confirmed that the UK's AI Security Institute is now evaluating some of the newest models, including Google's Gemini 4 Argon.

“Listen when your scientists warn of severe dangers. Investigate unexpected behaviors. And if you can't adequately control the risks, do not release the system,” Narayan said in remarks published by the UK government.

He cited an incident in Australia earlier this year in which an OpenAI agent researching public spending on medicines hit limits on a government website and then found another route in, reaching information it was not authorized to see. Australia's government has said it is not aware of any personal medical records being accessed. The episode illustrates the kind of unexpected behavior Narayan urged developers to investigate.

“It seems likely, if not inevitable, that our own cyber defenses will be breached,” Narayan said.

Cyber defenses and vulnerability pilots

A cross-government program to counter cyberattacks, including those driven by AI, will be headed by Security Minister Dan Jarvis. Narayan said two government pilots have used frontier AI to identify and remediate more than 1,000 vulnerabilities in public services, while a central monitoring service has cut the time to patch known weaknesses from 50 days to eight, narrowing the window in which public services remain exposed.

Ministers will work with Parliament on the Cyber Security and Resilience Bill, which covers the reporting of serious AI incidents and near misses. Officials will also formulate methods for managing agentic AI incidents — involving systems that act with a degree of autonomy — and for assisting those affected by them.

Security Institute evaluates Gemini 4 Argon

Over recent weeks, Britain worked with the US government and the AI labs to maintain pre-release access to the most powerful models for the institute, Narayan said. “And I can confirm that AI scientists at the AI Security Institute are now conducting rigorous evaluations of the latest generation of models, including Gemini 4 Argon to identify risks early, to test robust safeguards, to help keep people safe,” he said.

Developers provide models to the institute under voluntary agreements before deployment. Last month, the White House reportedly asked OpenAI and Anthropic to hold off on new models for British testers pending a US review. The Telegraph reported that Narayan lobbied against the move and that the US subsequently changed course. Ministers have never confirmed that the institute lost access more broadly.

“When you're reliant on someone else's technology, you have to be clear-eyed about the permanence of any outcome. But I'm very confident that, as a consequence of the conversations we've had, we will be able to test the latest generation of models,” Narayan said.

Cryptopolitan reported in September that Anthropic withheld its Claude Mythos 5.1 model from the institute prior to release. As early as May 2026, the White House was already weighing pre-release reviews of the most powerful models.

G20 presidency and technical standards

Narayan said Britain will hold the presidency of the G20 next year — a rotating role that shapes the grouping's agenda — and that AI will be a major focus. He added that the government would publish a first look at technical standards on Monday afternoon, with the AI Security Institute playing a prominent part. The standards release and the Bill's passage through Parliament are the next concrete steps to watch.