Anthropic Expands Project Glasswing to Over 15 Countries Anthropic is rapidly widening access to its powerful cybersecurity AI while trying to reassure governments and industry that the technology will be used to protect, not destabilize, critical infrastructure.

In early April, Anthropic launched Project Glasswing, a “collaborative effort to secure the world’s most important software,” and gave “roughly 50 initial partners” access to its Claude Mythos Preview model to scan codebases for vulnerabilities. Those early partners have already uncovered “more than 10,000 high- or critical-severity security flaws,” according to the company.

Following those results, Anthropic announced on June 2 that it is “giving Claude Mythos Preview to around 150 more organizations.” The expanded cohort spans “more than 15 countries” and is explicitly aimed at sectors that were “not well represented” at first — including power, water, healthcare, communications, and hardware.

From Anthropic’s perspective, the expansion is both a scale-up and a stress test for governance. The company says each new participant must meet security requirements before gaining access, and it frames the move as preparation for “cheap, fast AI models with powerful cyber capabilities” that are “around the corner.” Its long‑term goal, it says, is “for AI to make all software more secure” and to help the industry adjust as AI reshapes cybersecurity assumptions.

Human observers emphasize the geopolitical and infrastructure stakes. TechCrunch reports that Anthropic is “scal[ing] Claude Mythos to critical infrastructure in 15+ countries,” targeting systems where “a cyberattack could affect 100 million people” and carry “important ramifications for both global and national security.” The Verge highlights that organizations across power, water, and healthcare will now use the model “to find security vulnerabilities,” broadening the initiative beyond its initial, more limited cohort.

Together, the AI maker and outside analysts portray Project Glasswing’s next phase as a high‑stakes experiment: deploying frontier security tools to safeguard the very infrastructure that could be most at risk if such tools are misused.

Continue reading https://foxvector.com/stories/019e92fb-95e1-3104-7218-23b8a4f10e45