A coalition of five Western intelligence agencies, known as the Five Eyes alliance, warned Monday that artificial intelligence is accelerating cyber threats at a pace that could leave governments and businesses with mere months to prepare rather than years.
The joint statement from the United States, United Kingdom, Canada, Australia and New Zealand said AI's impact on both defensive and offensive cybersecurity capabilities is already underway. "While AI will help us improve cyber defence over time, it also accelerates the speed, scale, and sophistication of cyber threats," the countries wrote. "AI is not a future consideration — it is already here."
What the Right Is Saying
Republican lawmakers and free-market technology advocates argued that heavy-handed regulation would cede AI development leadership to China while failing to address actual security threats. They noted that the administration's recent directive prompting Anthropic to pull its newest models sparked backlash from researchers who said such interventions could chill innovation.
House Speaker Mike Johnson (R-La.) has championed what he calls a "pro-innovation" approach to AI governance. His office released a statement saying, "We can address cybersecurity risks without strangling the American tech sector with red tape that our adversaries will simply circumvent." The Heritage Foundation argued that mandatory pre-release testing would give competitors including China access to proprietary information about model capabilities.
Conservative technology commentators emphasized that AI also provides powerful defensive tools. They pointed to the Five Eyes statement's acknowledgment that "AI offers powerful tools to strengthen defence" as evidence that the private sector should be empowered to develop security solutions without government mandates. The American Enterprise Institute cautioned against policies that would push AI development offshore.
What the Left Is Saying
Democratic lawmakers and progressive technology policy advocates said the Five Eyes warning underscores the need for mandatory safety testing of advanced AI models before public release. They pointed to Anthropic's decision earlier this year to withhold its Mythos model, which the company said could identify decades-old software vulnerabilities that bad actors could exploit.
Senator Elizabeth Warren (D-Mass.) has called for legislation requiring pre-release security reviews of frontier AI models. "Voluntary frameworks haven't worked in financial services, they haven't worked in healthcare, and they won't work here," her office stated. The Progressive Policy Institute argued that the administration's recent executive order on voluntary testing represents a step backward from meaningful oversight.
Civil liberties groups aligned with progressive causes said the warning validates concerns raised during Congressional hearings about AI-enabled threats to critical infrastructure, including power grids, water systems and financial networks. They urged the administration to prioritize defensive capabilities over industry-friendly regulatory approaches.
What the Numbers Show
The Five Eyes joint statement did not include specific statistics on cyberattack frequency or AI-related incidents. However, the agencies highlighted that AI reduces the time between vulnerability discovery and exploitation, a metric cybersecurity firms have estimated is shrinking from weeks to days or hours with advanced AI tools.
Anthropic has stated its Mythos model can analyze decades of existing vulnerabilities across millions of code repositories in hours rather than months. The company said this capability could help organizations patch systems faster but acknowledged it would also enable malicious actors to identify targets at unprecedented speed and scale.
The Trump administration's executive order from earlier this month established a 30-day voluntary pre-release testing period for AI labs, though critics note there is no enforcement mechanism for non-compliance. Anthropic's Fable and Mythos models were subsequently pulled following a separate administration directive that required additional review.
The Bottom Line
The Five Eyes warning represents the most significant coordinated international statement on immediate AI cybersecurity risks to date, signaling that intelligence agencies across Western allied nations view frontier AI development as a near-term national security concern rather than a distant theoretical threat. The joint position puts pressure on the Trump administration to demonstrate its voluntary testing framework can keep pace with rapidly evolving capabilities.
What happens next will likely depend on whether any major cyber incidents linked to AI-enabled tools occur in the coming months. If attacks materialize as described, it would probably shift the political calculus toward mandatory safety protocols. If not, industry advocates may successfully argue that existing approaches are sufficient. Either way, the debate over how to balance innovation with security is expected to intensify when Congress reconvenes.