Anthropic has made a significant move in the cybersecurity arena by widening the scope of its Project Glasswing. This initiative, driven by advanced artificial intelligence, now includes approximately 150 new organizations under the Claude Mythos Preview program.
Expanding Global Cybersecurity Efforts
The recent expansion comes after Anthropic’s intensive collaboration with existing partners, the broader security industry, and governmental bodies. Initially launched in April 2026, Project Glasswing began with about 50 partner organizations focusing on identifying security vulnerabilities within their codebases.
These early collaborations resulted in the detection of over 10,000 high-severity vulnerabilities, highlighting the efficacy of Mythos-class AI models in safeguarding enterprise systems. The newly added organizations, spanning over 15 countries, predominantly provide critical infrastructure services, enhancing the program’s global impact.
Broadening Industry Participation
The expansion now includes sectors previously underrepresented, such as power, water, healthcare, communications, and hardware. Many of the new partners are vendors and nonprofits that maintain essential codebases used widely across various organizations, including governmental bodies.
Anthropic’s risk assessment emphasizes the potential impact of cyberattacks on these organizations, stressing the importance of stringent security measures before granting access to Claude Mythos Preview.
Enhanced Capabilities and Future Outlook
Beyond detecting vulnerabilities, Claude Mythos Preview is being utilized for a broader range of defensive operations. These include automated patching, pre-release security checks, penetration testing simulations, and more. The AI also acts as a crucial early warning system for the industry to prepare for future threats.
Anthropic predicts that similar AI models may emerge from competitors within the next year, potentially lacking adequate safeguards against misuse. This underscores the urgency for current cyberdefenders to enhance their strategies.
To further extend its defensive capabilities, Anthropic offers tools on request to trusted security teams and has introduced Claude Security, aimed at organizations beyond the program. The company is also engaging with third parties to enhance open-source software vulnerability management.
Anthropic’s long-term vision is clear: to transition AI’s role in cybersecurity from merely identifying vulnerabilities to a comprehensive model that includes disclosure, patching, and deployment, ensuring a sustainable advantage for defenders.
