Expanding Project Glasswing Project Glasswing is expanding its collaborative effort to secure critical software by including approximately 150 new organizations from over 15 countries, focusing on industries like power, water, and healthcare. The initiative aims to leverage AI, specifically Claude Mythos Preview, to scan codebases for vulnerabilities and prepare the cybersecurity industry for the impact of powerful AI models.

• Project Glasswing is expanding to include around 150 new organizations in over 15 countries, focusing on critical infrastructure and vendors.
• Initial partners using Claude Mythos Preview found over 10,000 high- or critical-severity security flaws.
• The expansion includes industries like power, water, healthcare, communications, and hardware, many of which provide critical infrastructure.
• The project aims to make AI enhance software security and help the industry adapt to AI’s impact on cybersecurity assumptions.
• Future AI models may be released without safeguards, increasing the frequency and unpredictability of cyberattacks.
• The company plans to release Claude Security and offer tools to trusted security teams to help scan codebases and suggest patches.
• The bottleneck in cybersecurity is now verifying, disclosing, and patching vulnerabilities surfaced by advanced AI models.
• Efforts are underway to scale up vulnerability review and patching in open-source software.
• Safeguards against misuse of powerful AI cyber capabilities are still under development.
• Future expansions will prioritize essential infrastructure providers, critical open-source software maintainers, and safety testers. Continue reading https://www.anthropic.com/news/expanding-project-glasswing