Months, Not Years: Five Eyes Issues Urgent AI Cyber Warning as Industry Leaders Admit Existential Risks

Five Eyes Cyber Security Agencies – including the US, UK, Canada, Australia, and New Zealand – issued a joint statement Monday warning that AI cyber threats will exceed defensive capabilities within months, not years. The statement calls upon leaders to act quickly and correctly by utilizing AI technology for defense to protect against AI-led cyber attacks.

“AI is not a future consideration – it is already here,” the statement reads. “It lowers barriers for malicious actors and increases the speed and complexity of attacks, shrinking the window between vulnerability discovery and exploitation ever more quickly. At the same time, AI offers powerful tools to strengthen defence.”

Some leaders in the AI race have long warned about the potential incoming dangers from the exponential growth of this technology, including totalitarian control, propaganda, and models developing unstable or power-seeking behaviors under pressure. Dario Amodei, CEO of Anthropic and the creator of the popular AI model Claude, has noted that while some say, “No action is too extreme when the fate of humanity is at stake,” this kind of framing can lead to backlash.

In May, Anthropic Co-Founder Jack Clark said there is a “non-zero chance” of AI “killing everyone on the planet.” He stressed that it is important to state that the risk hasn’t gone away. Clark said it would be best if the development of AI technology was slowed to give humanity the opportunity to prepare for these potential harms. Clark said this will not happen because of the commercial and geopolitical rivalries that are fueling the AI development race.

Elon Musk, CEO of X and the AI model Grok, said last year there is a 20% of annihilation from AI when he appeared on the Joe Rogan Experience. Musk has also said that AI is more dangerous than nuclear weapons.

While these CEOs are acknowledging the risks and dangers of the technology, they continue to move forward with the rapid-speed development. There are two primary types of concerns that are outlined when it comes to dangers caused by AI. One is the risk that the AI starts acting on its own accord with misaligned goals as Musk and Clark referenced. The other is when malicious actors can utilize the technology to bypass current existing defense systems that are important for the way society functions. The latter is what the Five Eyes statement is concerned about imminently within months if leaders do not properly prepare.

The Five Eyes statement says that new vulnerabilities will emerge, including zero-day vulnerabilities, and breaches will occur as the technology continues to evolve. Zero-day vulnerabilities are those that are unknown by the developer until a breach has occurred, so it can cause significant harm depending on the size of the vulnerability. An analogy has been used comparing it to a home owner who is not aware that there is a secret camouflage door where intruders can enter. The secret door isn’t revealed until someone has already entered the home without permission and that person may start damaging things before the homeowner has an opportunity to install a lock to prevent further intrusion and damage.

“Cyber resilience is not an IT issue – it is central to operational continuity and market trust,” the statement says. “Leaders who act now will reduce exposure, strengthen resilience, and build confidence with customers, partners, and investors. Those who delay will face growing and avoidable risk.”

The Five Eyes Cyber Security Agencies are offering five practical actions in addition to utilizing AI technology for defense to protect against attacks. They suggest limiting “unnecessary system access and external connectivity” to reduce the attack surface.

They recommend prioritizing security updates and addressing unsupported systems that are considered easy targets. Critical system access should be limited. Lastly, leaders should “test response plans, train and prepare teams, and assume breaches will occur.”

In a January essay, Amodei wrote “we are considerably closer to real danger in 2026 than we were in 2023.” He believes this is part of a pendulum swing that resulted from the “least sensible voice” arising to the top, “often through sensationalized social media accounts,” that focused on inevitable destruction or doomerism in a “quasi-religious” manner.

Amodei says most people have been thinking about AI risks from a sober and analytical perspective for many years. He insists the AI development culture was focused on AI risk, but now the unsensible doomeristic voices have swung the pendulum in the direction of AI opportunism.

Amodei further describes the powerful AI models that could come as soon as 2027, but also acknowledged it is impossible to predict with certainty when this development stage will occur. He said this model will have skills and intelligence that exceeds the smartest and most capable humans in the world. The technology will have the ability to take hours to work on a larger, more thorough project and work at 10 to 100 times the speed of the most efficient humans. He said this could be described as a “country of geniuses in a data center.”

There are a number of risks associated with this technology that Amodei laid out. He would be concerned about autonomy and misaligned goals, rogue actors, the ability to seize power, economic disruption, and radical destabilization as a result of the exponential technology improvements and rapid productivity.

Despite all of these clear and potential risks associated with AI technology, Amodei said many US policymakers “deny the existence of any AI risks” while they continue to debate about the “usual tired old hot-button issues.” He said, “Humanity needs to wake up, and this essay is an attempt – a possibly futile one, but it’s worth trying – to jolt people awake.”

These are the words of a CEO of one of the companies on the front lines of AI development. Amodei stresses that he is not communicating certainty or even likelihood, but said there is a responsibility for AI companies and governments to act sooner rather than later to protect the public against these potential harms.

“To be clear, I believe if we act decisively and carefully, the risks can be overcome—I would even say our odds are good,” Amodei wrote. “And there’s a hugely better world on the other side of it. But we need to understand that this is a serious civilizational challenge.”