The Five Eyes intelligence alliance issued a warning on 23 Jun 2026 that artificial intelligence could be used by criminals to execute cyberattacks within weeks or months, according to a joint statement from cyber chiefs of the British, US, Australian, New Zealand, and Canadian agencies. The alliance cited concerns that advanced AI models will soon be capable of identifying software vulnerabilities and breaching digital security systems protecting major businesses and private data. The warning follows US restrictions on AI group Anthropic's release of its Fable 5 and Mythos 5 models outside the United States due to security concerns, and assessments by the UK's AI Security Institute that current frontier AI models can conduct cyberattacks from start to finish.
Five Eyes Alliance Warns AI Timeline for Cyber Threats Is Months Not Years
The Five Eyes spy alliance stated in its joint announcement: "We are united in our call to action: the evolving landscape of artificial intelligence is rapidly transforming cyber-risk, and we must act swiftly to remain ahead. While AI will help us improve cyberdefence over time, it also accelerates the speed, scale and sophistication of cyberthreats."
The alliance added: "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."
The agencies urged business leaders to "prioritise foundational cybersecurity practices and controls," warning: "Those that do not will face growing operational and strategic disadvantage."
Anthropic Barred from Releasing Fable 5 and Mythos 5 Models Outside US
Donald Trump barred AI group Anthropic from unveiling its new models, Fable 5 and Mythos 5, outside of the US over security concerns. Anthropic previously stated Mythos Preview was too dangerous to release because it could easily find serious flaws in software. Fable and Mythos 5 are commercial versions with guardrails on them.
Anthropic created an industry group, made up of tech companies including Apple and Microsoft, to probe Mythos so they can fix any issues.
UK AI Security Institute Assesses Mythos Preview Cyberattack Capabilities
The UK's AI Security Institute (AISI) called Mythos Preview a "step-up" in AI that could carry out a cyberattack from start to finish. The institute stated the new ChatGPT-5.5 model "reaches a similar level of performance on our cyber-evaluations" as Mythos.
Adam Beaumont, AISI's interim director, stated: "Two years ago, that finding probably would have sounded quite wild. I think today it sounds sobering. In two more years, it will be routine."
FAQ
What did the Five Eyes alliance warn about AI on 23 Jun 2026?
The Five Eyes intelligence alliance warned that artificial intelligence could be used by criminals to execute cyberattacks within weeks or months, stating that the timeline for frontier AI models transforming cyber-capabilities is "months" not "years."
Why did the US bar Anthropic from releasing Fable 5 and Mythos 5 outside the United States?
Donald Trump barred Anthropic from releasing Fable 5 and Mythos 5 models outside the US over security concerns, following Anthropic's own assessment that Mythos Preview was too dangerous to release due to its ability to find serious software flaws.
What did the UK AI Security Institute say about Mythos Preview's cyberattack capabilities?
The UK's AI Security Institute called Mythos Preview a "step-up" in AI that could carry out a cyberattack from start to finish, and stated that ChatGPT-5.5 reaches a similar level of performance on cyber-evaluations as Mythos.
