en.Wedoany.com Reported - Anthropic has expanded its cybersecurity initiative Project Glasswing, built on the Claude Mythos Preview model, adding approximately 150 organizations just weeks after the program's launch, including initial partners, security firms, open-source maintainers, and government agencies.

Anthropic noted that organizations joining the program must meet specific security requirements before gaining access. This expansion extends the initiative's reach to organizations in more than 15 countries, covering industries such as healthcare, energy, telecommunications, technology, and other infrastructure operators. The company stated it plans to further broaden the program's geographic coverage over time.

"In critical infrastructure, a single exposed vulnerability can impact far more than just one company. The expansion of Glasswing gives more vital sectors the opportunity to prepare for the next phase of cybersecurity—where vulnerabilities are discovered faster, and reducing real-world risk is key," said Joe Saunders, CEO of RunSafe Security.

Project Glasswing launched in April, initially with approximately 50 organizations using Claude Mythos Preview to scan software vulnerabilities. According to the company, the program has identified over 10,000 high-severity and critical flaws since its launch. The initiative has sparked discussions within the software industry and government circles, helping shape the program's expansion goals and objectives.

Anthropic wrote: "Mythos Preview continues a long-term trend we have been warning about for some time: In the next 6 to 12 months, we expect many other AI companies to have Mythos-level models, and they may release these models without safeguards to prevent misuse. In that scenario, cyberattacks could occur more frequently and in more unpredictable forms. Cyber defenders must adapt to keep pace."

According to Anthropic, vulnerability discovery is no longer a major bottleneck, but security teams still need to validate findings, disclose issues, develop patches, and deploy updates. Jim Sherlock, Vice President and Head of AI and Cybersecurity R&D at ProCircular, noted: "My advice is to spend about five minutes processing Anthropic's latest announcement, then immediately get back to your own patch cycle, because that's where companies will take the hit. Expect the next wave of security advisories to come from your vendors, in massive volumes and at speeds faster than your change window can handle." He added: "A patch pipeline that cannot handle the influx of advisories and vulnerabilities will only become a massive backlog of good intentions."

In addition to expanding the program, Anthropic also plans to make the vulnerability discovery tools developed specifically for Project Glasswing available to trusted security teams upon request. This follows last week's announcement that Anthropic plans to release its Mythos-level model to all customers once additional cybersecurity safeguards are in place. The company concluded: "We are working as quickly as possible to safely and broadly release Mythos-level capabilities. To do this, we need highly robust safeguards to prevent misuse of the model's cyber capabilities—safeguards that we are still developing."

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com