Anthropic expands AI cybersecurity programme for critical infrastructure
Critical infrastructure providers join Anthropic to strengthen cybersecurity and software resilience.
AI company Anthropic has announced a major expansion of Project Glasswing, an initiative aimed at strengthening the security of critical software through AI-assisted vulnerability detection.
After initially providing access to around 50 organisations, the programme will expand to approximately 150 additional partners across more than 15 countries.
Project Glasswing provides selected organisations with access to Claude Mythos Preview, Anthropic’s cybersecurity-focused AI model. According to Anthropic, participating organisations have identified more than 10,000 high- and critical-severity software vulnerabilities through the programme.
The newly added participants include operators and vendors across critical infrastructure sectors such as power, water, healthcare, communications and hardware manufacturing.
Anthropic argues that increasingly capable AI systems could significantly reshape cybersecurity, creating both new defensive opportunities and new risks. The company says future AI models may enable defenders to identify, analyse and remediate vulnerabilities at greater scale, while also potentially enhancing the capabilities available to malicious actors.
Project Glasswing is intended to help critical organisations adapt before such capabilities become widely accessible.
Alongside the expansion, Anthropic said it plans to provide additional cybersecurity tools, support vulnerability remediation efforts and work with industry, governments and open-source software maintainers to strengthen cyber resilience.
Why does it matter?
The expansion of Project Glasswing highlights the growing role of AI in cybersecurity, particularly in vulnerability discovery and software security testing. As critical infrastructure operators face increasingly sophisticated cyber threats, AI-assisted tools may help identify and address security weaknesses more quickly.
At the same time, the initiative reflects broader concerns that advances in AI could benefit both defenders and attackers, increasing the importance of responsible deployment, coordinated security research and resilience planning across critical sectors.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!