In a stark intervention that underscores the escalating anxiety around artificial intelligence, the co-founder of AI safety firm Anthropic has issued a direct warning to the British Prime Minister: cede control of the technology to machines and you risk a future that no human would choose.
Dr. Samuel Hammond, a former Google Brain researcher turned AI ethicist, met with Number 10 officials yesterday to deliver what he called a ‘wake-up call to the Whitehall machine’. His message was simple: the race to build ever-more powerful AI systems is currently being run without guardrails, and the consequence could be a catastrophic loss of human agency.
‘We are building intelligence that will one day surpass our own, and unless we embed human values and control at the core of these systems, we are essentially handing the steering wheel to a driver who doesn’t know where the brakes are,’ Hammond told The Daily Chronicle after the meeting. ‘The UK has a chance to lead the world in safe AI development, but that requires legislation, not just lip service.’
The urgency of Hammond’s visit reflects a growing split in the tech world. On one side are the ‘accelerationists’ who believe AI progress should proceed at full throttle, trusting that benefits will outweigh risks. On the other are safety advocates who warn of ‘alignment failure’ – where an AI pursues goals misaligned with human wellbeing, potentially with catastrophic consequences.
Hammond’s company, Anthropic, has positioned itself as a responsible alternative to rivals like OpenAI and DeepMind, emphasising ‘constitutional AI’ – systems trained to adhere to a set of ethical principles. But he admitted that voluntary measures are not enough. ‘Companies cannot police themselves. We need international standards, and the UK, with its history of scientific rigour and regulatory leadership, should be at the forefront.’
The Prime Minister’s office responded cautiously, reiterating the government’s commitment to ‘safe and responsible AI innovation’ without committing to specific legislative timelines. A Downing Street spokesperson said: ‘We welcome engagement from experts like Dr. Hammond and are carefully considering the regulatory framework needed to harness AI’s benefits while protecting the public.’
But critics argue that the government’s approach remains too cosy with Big Tech. Last month, the UK hosted a global AI safety summit but stopped short of endorsing binding rules on development or deployment. Instead, it secured a voluntary ‘pledge’ from leading companies to test frontier models for safety – a move dismissed by campaigners as ‘toothless’.
Hammond’s intervention comes as new polling shows that 67% of Britons believe AI should be subject to strict government oversight, and 54% are worried about losing control over decisions that affect their lives. ‘The public gets it,’ he said. ‘They see their data being harvested, their jobs automated, and now they worry that the very fabric of democracy could be warped by algorithms they don’t understand.’
The co-founder highlighted three immediate risks: AI-enabled disinformation that could undermine elections, autonomous weapons that make life-and-death decisions, and economic inequality amplified by large-scale job displacement. ‘Each of these is a clear and present danger if we treat AI as just another product. It is infrastructure – like electricity or the internet – and it needs public governance.’
Yet Hammond also offered a note of optimism. ‘The UK can be the Silicon Valley of safe AI. We have world-leading universities, a legal system built on precedent and ethics, and a public that demands responsibility. But that window is closing. Every month of delay is a month where the technology grows faster than our ability to control it.’
The room fell silent when he delivered his closing line: ‘We are the first generation to decide whether AI serves humanity, or humanity serves AI. The next election might be the last one that humans truly decide.’
