Cutting-edge artificial intelligence technology is on the brink of enhancing offensive hacking capabilities, prompting a call to action against the looming threat, as stated by officials from U.S., British, Canadian, Australian, and New Zealand cybersecurity agencies on Monday.
Members of the intelligence alliance, known as the Five Eyes, emphasized in a recent statement that advanced AI models are expected to surpass current industry standards, leading to a significant transformation in both offensive and defensive cyber capabilities in a matter of months, not years.
Among the officials endorsing this statement was Rajiv Gupta, the head of the Canadian Centre for Cyber Security, a key part of Canada’s national cryptologic agency, the Communications Security Establishment Canada (CSE). The cyber center highlighted frontier AI models as the latest, most capable AI models that are increasingly accessible and pose growing concerns regarding potential threats.
A cybersecurity expert from Cincinnati, Ohio, John Bruggeman, explained the substantial gap between powerful frontier AI models and their less advanced counterparts using an analogy, likening the former to the vast U.S. Library of Congress in terms of size and potency, compared to a standard city library.
The statement issued by the Five Eyes members on Monday provided limited specifics and mainly reiterated fundamental cybersecurity advice, such as promptly updating flawed software and minimizing unnecessary online system exposure. Additionally, the officials recommended leveraging AI to fortify defense mechanisms, enabling quicker identification of vulnerabilities and more rapid incident responses.
The warning underscored growing apprehensions among officials regarding AI models like Anthropic’s Mythos or OpenAI’s GPT-5.5-Cyber, which reportedly empower users to execute intricate and potentially catastrophic cyberattacks efficiently. Notably, Anthropic recently disabled a version of Mythos following a U.S. government directive suspending access to the models for foreign nationals due to alleged national security risks.
In a related move, the U.S. cyber defense agency CISA, a signatory of the Monday statement, tightened deadlines for government officials to address critical digital vulnerabilities in their networks to three days, citing the increasing AI-related threats.