June 23, 2026

Five Eyes warns new AI models pose urgent cyber risk

The “Five Eyes” intelligence alliance, comprising officials from the U.S., the UK, Canada, Australia, and New Zealand, has issued a warning that advanced artificial intelligence models, such as Anthropic’s “Mythos” and OpenAI’s “GPT-5.5-Cyber,” pose an urgent cyber risk, with the potential to significantly enhance offensive hacking capabilities in just months. Their recent statement emphasized the need for immediate action to address these threats by adhering to established cybersecurity practices, including rapidly patching vulnerabilities and minimizing system exposure, while also advocating for the use of AI to strengthen defensive measures. This warning follows a recent incident where Anthropic was ordered to restrict access to its AI model due to national security concerns, highlighting the escalating urgency among U.S. agencies to address AI-related vulnerabilities.

CISA: CISA is the U.S. Cybersecurity and Infrastructure Security Agency responsible for protecting critical infrastructure. It co-signed the Five Eyes statement urging immediate action against AI-enhanced cyber risks. The agency recently shortened government deadlines for addressing serious digital vulnerabilities to three days in response to emerging AI threats.
Mythos: Mythos is a frontier AI model developed by Anthropic. It was specifically cited in the intelligence alliance statement as capable of allowing users to execute potentially devastating cyber operations. Access restrictions were imposed on the model following U.S. government intervention earlier in the month.
OpenAI: OpenAI is an AI company that builds and deploys large-scale language models. It developed the GPT-5.5-Cyber model named in the Five Eyes warning for its potential to accelerate offensive hacking. The company’s models are among those officials say are exceeding prior expectations in cyber applications.
Anthropic: Anthropic is an AI company that develops advanced frontier models. It created the Mythos model referenced in the Five Eyes warning as an example of technology that could enable complex hacks. Earlier in June, the company was ordered by the U.S. government to suspend access to a version of the model for foreign nationals over national security concerns.
Five Eyes: The Five Eyes is an intelligence-sharing alliance of the United States, United Kingdom, Canada, Australia, and New Zealand. It issued a joint statement on June 22 warning that frontier AI models are expected to transform offensive cyber capabilities within months rather than years. The alliance called for urgent defensive measures including faster patching and greater use of AI for protection.
GPT-5.5-Cyber: GPT-5.5-Cyber is a specialized AI model from OpenAI focused on cyber-related capabilities. It was highlighted by the Five Eyes alliance as an example of technology poised to fundamentally alter both offensive and defensive cyber operations. The model illustrates the rapid timeline officials described for AI-driven threats.

AI Cyber Risks: Intelligence officials across allied nations are publicly emphasizing that frontier AI models are advancing cyber offense capabilities on a timeline of months rather than years.
Defensive Measures: The Five Eyes statement reinforces longstanding advice to prioritize rapid software patching and limit unnecessary system exposure while also encouraging the use of AI to bolster defenses.
Government Response: U.S. agencies have tightened controls on AI model access and accelerated vulnerability remediation timelines due to concerns over national security and AI-enabled hacking.

AI tools built by Emerald Force

Built and supported by Emerald Force.

You might also like