The UK government has left the door open to tougher AI regulation after the companies building the world’s most powerful models called for slower development, fearing safety has fallen behind.
City AM understands ministers believe AI is already covered by laws spanning data protection, competition, product safety and online safety, but accept those rules may no longer be sufficient as models become more capable.
Their position just stops short of committing to new legislation but marks a potentially significant shift as OpenAI itself urged Britain to impose binding rules on frontier developers.
“We take the risks from advanced AI extremely seriously and recognise the importance of ensuring our approach keeps pace with a rapidly evolving technology, a government spokesperson told City AM. “We should not assume that the existing framework will always be sufficient as AI capabilities develop.”
Any new measures would be proportionate and targeted at specific risks while seeking to minimise the impact on growth and innovation, City AM understands.
The government said it was working with leading AI companies and international partners, while its AI Security Institute was building a “rigorous, scientific understanding” of the risks posed by the most advanced systems.
The intervention comes after Anthropic chief Dario Amodei called for developers to “slow the pace” of improvements to frontier AI, warning companies needed to give safety work time to catch up. OpenAI boss Sam Altman and Elon Musk subsequently backed calls for slower development.
This comes as business secretary Jonathan Reynolds sought to temper the increasingly worrying warnings on Tuesday, saying ministers should “never be complacent”, but should also avoid becoming “hyperbolic”.
“There are some tremendous upsides for people as well, whether it’s public services, healthcare, or the contribution to the economy,” he said.
Reynolds also played down proposals for a mandatory AI “kill switch”, questioning how such a system could work when the most advanced models are largely developed in the US and China.
Mounting industry pressure #
OpenAI’s European policy chief Tom Duff Gordon told Politico the UK should “act now to set up a stronger set of safety rules”, arguing the government should exploit the “political window” created by the latest warnings.
The ChatGPT maker said it would support mandatory requirements for the small number of companies developing frontier models, including independent testing and reporting of safety incidents.
Labour pledged before entering government to introduce binding regulation for companies developing the most powerful AI models, but has instead relied heavily on existing regulators and voluntary agreements with developers.
This approach was challenged this week by parliament’s Joint Committee on Human Rights, which called for a new AI Bill and a statutory independent regulator.
Committee chair Alex Sobel said the existing system was not “fit for purpose”, warning regulators cannot currently prevent the release of a model they consider unacceptably risky.
Secretary of state Louise Haigh is expected to add to the pressure on Tuesday, saying ministers must “heed the warnings” from those developing the technology and work internationally to protect public safety and national security.
The debate has also widened the divide with Washington. Donald Trump attacked attempts to constrain American AI companies on Monday, arguing any slowdown could hand an advantage to China.
“There is a SICK conspiracy going on against AI and data centres,” the US president wrote on Truth Social. “WHOEVER WINS AI, WINS!”