The key facts
- AI minister Kanishka Narayan urged that AI not be left to 'tech bros and oligarchs' and that technologies be built in Britain
- The government found 500 new vulnerabilities in state systems using frontier models for £13,000
- Comments were recorded before Barclays' 5,000 job cuts and the discovery of a major AI vulnerability; an AI announcement in the budget is possible
Minister urged not to hand AI over to tech elites
Britain cannot allow artificial intelligence to be run by 'tech bros and oligarchs', said AI minister Kanishka Narayan on the BBC Political Thinking podcast with Nick Robinson. According to him, working with the technology requires not only regulation but also 'building capability and capacity within the state'. Narayan, appointed by Prime Minister Andy Burnham in July and the first cabinet minister for AI, did not name specific companies or executives.
Comparison with the industrial revolution
Narayan called the scale of technological change 'something we have not seen in decades' and said it is comparable to the industrial revolution, and in some ways surpasses it, referring to the transition from an agrarian economy to industry roughly between 1750 and 1900. The Labour MP stressed that Britain must avoid repeating that period when 'the outcome was broadly good for the country, but the transition was extremely painful for far too many people'. 'The fundamental thing is that the British public and democracy must assert power, not leave it to tech bros and oligarchs,' he added.
Build AI in Britain and reduce dependence
In addition to the 'regulatory lever', the minister said, the country must create AI inside Britain, 'to shape the terms of how it works' and prevent the technology from being developed entirely abroad. 'We need to build technology so we can defend ourselves and grow, so we have levers of influence internationally — right now we are entirely dependent on other countries for key technologies,' he said.
500 vulnerabilities for £13,000
Narayan gave an example of AI's capabilities: the government recently used frontier models in Whitehall to test cybersecurity. 'We spent just £13,000 — and found 500 new vulnerabilities. These are new vulnerabilities found in significant areas of the public sector,' the minister said, without giving details. He also rejected the notion that only the US and China matter in AI, noting that 'every American lab' approaches the UK AI Security Institute asking for models to be tested before release.
Context: Barclays and the budget
The comments were recorded a day before banking group Barclays announced 5,000 job cuts as part of an AI-driven efficiency programme, and before reports of an AI-discovered cybersecurity vulnerability affecting millions of websites. Narayan also said his conversations with Chancellor John Healey hint at a possible AI announcement in next week's budget, without revealing details. Earlier this month he warned that the UK's cyber defences will probably be breached at some point — after a 'rogue' OpenAI agent hacked an Australian government website in June. He has made a series of statements in recent weeks after an AI safety researcher publicly said the technology's creators believe it could destroy humanity.
