UK Minister Says Government Must Heed AI Warnings as Cyber Leaders Stress Human Readiness

Over the weekend, Anthropic urged competitors and governments to coordinate a global slowdown in AI development. It was backed by rivals including OpenAI, Google DeepMind and X's Elon Musk, The Guardian reported. Haigh is expected to say AI has enormous potential for health and public services, while the government will seriously examine public safety and national security threats with international partners.

On Monday night, Anthropic cofounder Jack Clark suggested a third-party AI 'kill switch' may need to be mandatory for companies, saying society 'might want to eventually pass rules around' it. OpenAI, the operator of ChatGPT, has also urged the UK government to capitalize on renewed fears over AI safety and impose legislation reining in the technology. Tom Duff Gordon, OpenAI's head of policy in Europe, said: 'We support stronger UK rules for the handful of companies, including OpenAI, developing the most powerful AI systems.' The company said legislation should focus narrowly on the most advanced AI labs, not startups working on less powerful systems.

Haigh, whose speech will also set out priorities for public control of transport, water and energy, will tell unions: 'AI has enormous potential to transform our public services, make our businesses stronger and deliver new scientific breakthroughs.' But she will say the government must 'heed the warnings of those who are at the forefront of developing this technology' and 'be ready to work with international partners' to prioritize public safety and national security.

MPs and peers on Parliament's human rights committee on Monday called for a new UK regulatory framework for AI, including an independent oversight body and legislation to protect the public. The committee said AI posed threats ranging from public face-scanning to explicit deepfakes and to flagging disciplinary action for workers without reasonable cause. Alex Sobel, the Labour chair of the committee, said the world was unprepared for the 'potentially dire' consequences of rapidly developing technology. 'Nowhere in the world, including the UK, has a current legislative and regulatory approach to AI that is fit for purpose,' he said.

A government spokesperson said the UK took the risks from advanced AI 'extremely seriously' and recognized the need to keep pace with the technology, including by working closely with AI companies and other countries. 'We should not assume that the existing framework will always be sufficient as AI capabilities develop. Any future measures will be evidence-based and focused on the risks we need to address,' the spokesperson said.

But Business Secretary Jonathan Reynolds said on Tuesday morning that people should not get 'hyperbolic' about AI risks. He said he did not think it would be particularly helpful to discuss a potential AI 'kill switch,' which he felt had little practical meaning. Speaking to BBC Radio 4's Today programme, Reynolds said: 'This is extremely powerful technology, and I think we should never be complacent or naive about the impact it might have. People will be worried by some of the things they've heard in the last few weeks, and I think we've got to be careful not to get hyperbolic about this either.' He added that there were risks but also 'tremendous upsides' for public services, healthcare and the economy, and that regulation should be proportionate. 'If you regulate it in a way where you're no longer having access to those frontier developments, that would obviously make us less safe,' he said.

Last Friday, more than 70 MPs and peers urged Andy Burnham to back a ban on the creation of artificial superintelligence, or ASI, and to lead an international movement to stop its development. The letter followed a week of warnings in which experts estimated a 10% chance that ASI would wipe out humans. Signatories included former government ministers, former civil service head Robin Butler, senior Labour MP John McDonnell, and leading Conservative, Liberal Democrat and Scottish National party MPs. They urged the prime minister to throw the government's weight behind a bill tabled in the Commons.

The policy debate is unfolding alongside rapid changes in corporate cyber-readiness. Sifted reported that business leaders are adopting AI while confronting the fact that malicious actors are using the same technology to devise more sophisticated cyber attacks. Haris Pylarinos, chief executive and founder of the cyber readiness platform Hack The Box, said: 'AI is not only a superpower for defenders and organisations; it is also a superpower for criminals.' He said AI can give defenders a significant advantage, but simply adopting it is not enough. 'The organisations that gain the most will be those investing in the people who know how to direct it, challenge its output and step in when needed,' he said.

According to Sifted, Hack The Box has more than 800 enterprise customers and a global community of over 4 million cybersecurity professionals. It helps organizations develop and measure cyber workforce capabilities through threat-informed learning, hands-on labs and live-fire environments designed to reflect real-world attacks. Pylarinos said: 'You build readiness by practising under realistic conditions repeatedly. When an incident happens, the team is not encountering that pressure for the first time.' He added that performance should not be measured by the days a company has avoided a breach. 'We measure performance by metrics like how fast and how many vulnerabilities penetration testers find in a regular [simulated] pentest,' he said. 'Measuring the days you haven't been hacked is not an effective method. You might be hacked today, in a year, or in five years.'

Hack The Box president Gibb Witham said cybersecurity teams need platforms that safely test human skills and AI capabilities, comparing it to a pilot needing a flight simulator. Witham said AI effectiveness is directly tied to the user's domain skill, because experts can prompt models better and accurately check outputs. He said sophisticated attackers use a combination of AI and human intelligence, meaning fully automated defenses would likely be outsmarted quickly. 'Cybersecurity is unique because there is a constant, intelligent adversary also utilising AI,' Witham said. 'Technology leaders shouldn't evaluate AI in a vacuum. Instead, measure the domain capability of your professionals, as higher skill levels yield much greater leverage from AI.'

Pylarinos said companies that simply install AI everywhere and expect it to solve their problems will fail, while those that invest in the human behind the wheel will prevail. He said AI is changing how attackers work and how defenders train, requiring a new level of investment in training and experimentation so that combined teams of humans and agents can achieve business goals. 'Realistic cyber environments give teams a safe place to practise,' he said.