‘Act now’: AI models capable of attacks on governments months away, Five Eyes warn
Powerful AI models capable of devastating cyber attacks on both governments and businesses are just months away, intelligence agencies for the coveted Five Eyes has warned in a rare joint statement.
The cyber security agencies from Australia, the US, UK, New Zealand and Canada, urged leaders to “act now” in a rare public intervention after the Trump administration decided to block “foreign nationals” from using the Fable AI model built by tech-heavyweight Anthropic.
The statement, which was issued on Monday, argued that while AI “would help us improve cyber defence over time, it also accelerates the speed, scale and sophistication of cyber threats”.
The group said in a statement: “Frontier AI models are anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities. The timeline is not years, it is months.
“In this environment, cyber resilience is integral to advancing business continuity, market confidence, and long-term value.”
The Five Eyes agencies, which are part of an intelligence alliance set up between the five countries after the second world war, went on to say that the fast growth in AI models capabilities showed the technology would lower barriers for bad actors to increase both the speed and complexity of their attacks.
Generative AI models are capable of looking for vulnerabilities in cyber security systems in a bid to exploit them as well as repair them.
The statement continued: “A whole-of-organisation and whole-of-society response is required. Cyber risk can no longer be treated as a purely technical issue. This is a core business risk and leadership responsibility.”
All eyes on AI
While the statement from Five Eyes failed to mention any specific AI company or model, eyes across the globe have swivelled to Anthropic’s advanced technology.
It’s latest invention, dubbed Fable 5, is a supposed more approachable version of Mythos, a powerful AI model released earlier this year which is capable of detecting cracks in cyber systems.
The Mythos model is only available to vetted organisations and companies amid concerns of it being exploited, with these fears ultimately leading the US government to suspend the use of both models by “foreign nationals” earlier this month, citing advice from national security authorities.
Industry figures have also warned that the UK is particularly exposed to bad actors and, if it fails to act, risks “sleep-walking” into their clutch.
Andy Ward, SVP International at Absolute Security, said: “This warning is just the beginning, attacks now move at AI speed, disruption moves at AI speed, so if your cyber defences don’t keep up, you’ve already lost.
“Our recent research found that 42 per cent of UK organisations still lack a formal cyber resilience strategy. Without robust AI-powered cyber resilience strategies and real-time visibility in place, the UK risks sleepwalking into deeper vulnerabilities.”
Despite fears of Anthropic’s advancing tech, the company is charging ahead with its IPO plans, after filing confidentially for a US stock market listing and advancing its rivalry with Open AI.
The Claude maker confirmed that it had submitted a draft registration statement on Form S-1 to the Securities and Exchange Commission at the start of the month, giving the company the option to go public once regulators complete their review.
The company did not disclose the size, timing or valuation of the offering.