AI Risks Existential Threat, Warns UK Foreign Secretary

Governments Urged to Act on AI Regulation Before Crisis

London, UK – Governments worldwide risk repeating the cautionary tales of the nuclear age if they delay establishing regulations for artificial intelligence, warned Yvette Cooper, the UK’s Foreign Secretary.

In an article published Monday, Cooper highlighted that while AI offers significant advancements, it also presents escalating risks as the technology becomes more potent and accessible.

“Last month, in Shenzhen, China, I witnessed the extraordinary promise of AI and robotics applied to life-saving healthcare,” Cooper wrote. “However, these same technologies are also alarmingly reshaping the future of warfare, crime, and social cohesion.”

Cooper asserted that managing the dangers posed by AI could become “the greatest security challenge of the next decade.” She argued for the necessity of international agreements on frontier technologies before a crisis compels action.

She drew a parallel between the current drive to develop AI systems and the early nuclear arms race, noting that global safety accords only materialized after nations experienced the destructive power of atomic weapons.

“Regarding nuclear technology, international agreement only emerged after the world witnessed the terrifying power of the new technology at Hiroshima,” Cooper stated. “We cannot afford to wait for an AI equivalent of Hiroshima before we act.”

Cooper called on Britain to leverage its diplomatic influence to convene the United States, China, and other leading AI nations to forge shared safety principles and standards. She pointed to the 2023 AI Safety Summit at Bletchley Park, where global leaders from 29 countries and the European Union convened to discuss the emerging risks of AI, as an example of the UK’s capacity to “rally the world on AI security.”

Cooper’s warning follows months of mounting concerns over how governments should oversee increasingly sophisticated AI systems.

In May, the UK’s AI Security Institute cautioned about rapid advancements in AI cybersecurity capabilities after OpenAI’s GPT-5.5 became the second model to successfully execute a simulated cyberattack without human intervention, following Anthropic’s Claude Mythos Preview.

Days later, the International Monetary Fund issued a warning that AI could “amplify” cyberattacks against the global financial system by reducing the expertise required to exploit vulnerabilities. The IMF urged policymakers to address cybersecurity as a financial stability issue, rather than solely a technical concern.

In June, President Donald Trump signed an executive order establishing a voluntary framework for reviewing advanced AI models prior to their release, expanding AI cybersecurity programs, and directing agencies to assess potential national security risks associated with frontier AI models.

That same month, calls for more stringent regulations also emerged from within the AI industry. Anthropic CEO Dario Amodei argued that transparency requirements are no longer sufficient and advocated for mandatory third-party testing of frontier models. This demand was followed by a U.S. government order for Anthropic to restrict access to Claude Fable 5 and Mythos 5 due to national security concerns, an order that was later rescinded in July.