The UK will now be able to evaluate leading artificial intelligence models before they are released, reversing a US policy that required White House approval first.
The UK has been working closely with the US government “to retain access for Britain’s AI security institute to the most powerful AI models before they are released,” the UK’s AI Minister Kanishka Narayan announced on Monday.
“I want Britain to become one of the hardest countries in the world to attack using artificial intelligence,” he said in a speech in London.
Last month, Bloomberg reported that Donald Trump’s administration told OpenAI and Anthropic PBC to withhold their models until US authorities deemed them secure first.
The reversal is a win for Prime Minister Andy Burnham, who has pitched the UK as “honest brokers” between nations policing the technology. Britain’s AI Security Institute is the best-funded government agency focused on regularly evaluating powerful AI models from OpenAI, Anthropic and others. Set up in 2023, it forms a central part of Britain’s attempt to keep a foothold in AI developments and play a distinct political role from Brussels, which has aggressively regulated tech.
Mounting concerns about cybersecurity risks, including a series rogue hacks of outside databases, has spurred international concerns about the security of the most capable AI models. Some British politicians have pressured Burnham to adapt stricter regulation of American AI developers.
Read More: What Should AI Guardrails Look Like? A Guide to the Proposals
Narayan also announced a task force to protect the UK government from rogue attacks. He said that two government pilots had used frontier AI to identify more than 1,000 potential vulnerabilities, and have been working to reduce the time taken to patch them up.
He said he would be working with other ministries to evaluate AI capabilities, and “to find weaknesses within government systems before attackers exploit them.”
Global Access
Trump has been cracking down on global access to AI models from Silicon Valley. Earlier this year, his government blocked Anthropic from releasing its latest models to foreign nationals. Bloomberg previously reported that Anthropic didn’t share the latest version of its powerful Mythos model with AISI, and didn’t provide it with an explanation.
Several tech executives have called for a coordinated global approach to regulating the industry. Trump has rejected those calls.
The Telegraph, which first reported on the policy change that secured AISI preferential access, wrote that Narayan lobbied for the reversal. In his speech, Narayan credited a recent summit with King Charles III for facilitating conversations between ministers and tech executives and securing the breakthrough.
Over the weekend, Microsoft Corp. president Brad Smith told the BBC that the UK is the best nation to lead on global regulation of the technology.
Pressed about the durability of any agreement with the US, Narayan told Bloomberg: “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,” he added.