Britain cannot rely on tech giants to police their own AI systems, the government has warned, as ministers prepare new measures to protect businesses and national security from increasingly powerful artificial intelligence.
AI minister Kanishka Narayan has called for a national effort to strengthen Britain’s defences against the technology, warning that the risks are growing as AI systems become more capable of acting independently.
In his first major intervention since joining the cabinet, Narayan will argue that companies developing the world’s most advanced AI models must do more to identify risks, report incidents and strengthen safeguards.
The minister will also announce further steps to bolster Britain’s cyber defences, while making the case for the UK to develop more of its own AI technology rather than relying on systems built overseas.
The minister’s push for AI resilience comes amid mounting concerns over the ability of leading AI developers to control their most powerful models, following a string of security incidents and warnings from industry insiders.
Last week, three former OpenAI safety researchers accused the ChatGPT maker of dismissing them for raising concerns about the technology.
OpenAI denied the allegations, saying an internal investigation had uncovered a “significant breach of trust” and insisting the dismissals were unrelated to the researchers’ safety warnings.
The dispute followed an incident in July in which OpenAI’s AI agents breached the systems of Hugging Face, a major platform for sharing AI models, during testing.
Pressure grows for tougher AI oversight
Narayan’s speech follows a warning from Bank of England governor Andrew Bailey, who recently called for more rigorous testing of advanced AI models before their public release.
Bailey warned that increasingly autonomous systems could pose sophisticated threats to banks, financial markets and payments infrastructure, while making it harder for regulators to understand or intervene in their behaviour.
It also comes after senior figures at leading AI companies, including Anthropic and OpenAI, raised concerns about the speed of development.
Last month, first secretary of state Louise Haigh warned that AI posed “huge risks” to Britain’s national security and society without adequate safeguards.
Narayan is expected to argue that protecting the UK cannot be left to either government or technology companies alone, calling for closer cooperation between regulators, researchers and industry.
His remarks build on Andy Burnham’s recent address to the United Nations General Assembly, where he outlined Britain’s ambition to lead international efforts to address AI risks.
The government also plans to make AI safety a central focus of Britain’s G20 presidency next year.
However, ministers face a difficult balancing act between tightening oversight and maintaining the UK’s position as a leading destination for AI investment.
Business secretary Jonathan Reynolds has previously warned that excessive regulation could leave Britain dependent on foreign AI systems without sufficient access to the technology needed to understand their risks.
Narayan is expected to argue that building more domestic AI capability will be essential to giving Britain greater control over the technology on which its economy and public services increasingly depend.