Five Eyes Agencies Warn Advanced AI Models Months Away

Urgent warning on AI-powered cyber threats

Cybersecurity agencies from the Five Eyes intelligence alliance have issued an unusual joint statement warning that artificial intelligence models capable of destabilizing governments and businesses will arrive within months, not years. The public intervention by signals agencies from Australia, the United States, the United Kingdom, New Zealand, and Canada calls on leaders to "act now" as AI fundamentally transforms both offensive and defensive cyber capabilities.

The warning follows the Trump administration's decision earlier in June to block foreign nationals from accessing Fable, an advanced AI model developed by Anthropic. The restriction, based on advice from national security authorities, applies to both Fable 5 and Mythos, Anthropic's earlier model that can detect vulnerabilities in cyber systems and remains available only to vetted organizations.

According to The Guardian, which first reported the statement issued late Monday Sydney time, the agencies emphasized that while AI will improve cyber defense over time, it simultaneously accelerates the speed, scale, and sophistication of cyber threats.

Why it matters

The joint statement represents a significant escalation in how intelligence agencies view AI risks. By publicly warning that frontier AI models will "exceed current industry expectations" within months, Five Eyes agencies are signaling that the technology has advanced faster than anticipated and that current security frameworks may be inadequate. The intervention also highlights growing tensions between AI development and national security, particularly as governments struggle to balance innovation with risk management.

Lowering barriers for malicious actors

The cybersecurity agencies stated that emerging AI models will lower barriers for bad actors while increasing the speed and complexity of attacks. They emphasized that cyber risk can no longer be treated as a purely technical issue, calling it "a core business risk and leadership responsibility" that requires responses from entire organizations and societies.

While the statement does not name specific AI models or companies, attention has focused on Anthropic's advanced tools. The company's Mythos model, released earlier this year, demonstrated capabilities for detecting system vulnerabilities that raised concerns about potential exploitation.

Beyond Anthropic's models

Olivia Shen, an expert in national security and AI at the University of Sydney's United States Studies Centre, cautioned that the focus on Anthropic's models may miss other developments. She noted that other states, actors, and companies—including China—could be developing equally advanced models that remain hidden from public view.

"We can only see what's been released, but there could be other models being developed by the likes of China, or other states and other actors and companies, that are just as advanced," Shen said.

In March, the Australian government signed Anthropic as the first company in its national AI plan, a non-binding memorandum of understanding requiring companies to share AI progress details with the government and promote safety. The government has adopted a light-touch regulatory approach aimed at capturing economic and productivity benefits from AI technology.

Details of the Five Eyes warning and the restrictions on Anthropic's models were first reported by The Guardian.