Anthropic Expands Access to ‘Mythos’ Cybersecurity AI Model Anthropic’s push to deploy its powerful “Mythos” cybersecurity AI worldwide is accelerating, even as it stirs political, strategic, and ethical debates about who controls such capabilities.

In early April, Anthropic launched Project Glasswing with roughly 50 initial partners using Claude Mythos Preview to scan code for weaknesses, a collaboration that the company says has already uncovered “more than 10,000 high- or critical-severity security flaws.” The first phase largely involved major tech companies and U.S.-aligned government entities.

Weeks of pressure from Europe followed. EU institutions learned Mythos had found vulnerabilities in software they relied on, while “no European institution could see the findings.” After a tense standoff, Anthropic agreed to give the EU cybersecurity agency ENISA access, making it the first EU body inside Glasswing and ending “one of the most visible flashpoints in the transatlantic AI relationship.” Parallel talks continued on a broader deal as “Anthropic offers EU access to Mythos” beyond the U.S. and UK.

On June 2, Anthropic formally announced it was “expanding Project Glasswing” to approximately 150 new organizations across more than 15 countries, with a focus on power, water, healthcare, communications, and hardware — sectors where “a successful attack on their codebase could be catastrophic” and “affect more than 100 million people.” TechCrunch described the move as scaling Claude Mythos “to critical infrastructure in 15+ countries,” targeting systems where a cyberattack could impact national security. The Verge similarly reported that Anthropic is giving Claude Mythos Preview “to around 150 more organizations” so they can use it “to find security vulnerabilities.”

At the same time, geopolitical and ethical concerns are sharpening. The Financial Times has reported both that “Anthropic to expand Mythos access to more than 15 countries” and that the NSA is using Mythos “for Cyber Attacks,” even as Anthropic remains “locked in [a] legal battle with [the] Pentagon over [its] Claude model.” European commentators warn that ENISA’s access underscores “broader European concerns about dependency on foreign AI infrastructure for critical security functions.”

Industry critics have gone further, questioning the ambition behind frontier labs. In a widely shared discussion amplified by AI pioneer Yann LeCun, investor Bill Gurley claimed that “Anthropic thinks it’s building God,” calling it “the ultimate level of narcissism and delusion of grandeur to think you can create God.” As Mythos moves from a small pilot to a core tool for global infrastructure and intelligence, those concerns about power, control, and intent are likely to intensify.

Continue reading https://foxvector.com/stories/019e9822-4339-0f71-7010-1c5295a1de97