Anthropic scales Claude Mythos to critical infrastructure in 15 countries
5 hours ago
- #AI Security
- #Software Vulnerabilities
- #Cybersecurity
- Anthropic expands Project Glasswing to about 150 organizations across 15+ countries to find critical software vulnerabilities using AI.
- Project Glasswing centers on Claude Mythos, Anthropic's most powerful AI model, which identifies zero-day vulnerabilities.
- Initial access was given to 50 partners including the U.S. government; expansion now includes critical infrastructure sectors like power, water, healthcare, communications, and hardware.
- The new partners are organizations where a successful attack could affect over 100 million people and impact global/national security.
- Expanded access includes organizations in U.S.-friendly countries such as Australia, Canada, France, Germany, Japan, India, and others.
- Reported organizations with access include Okta, Samsung, SK Hynix, SK Telecom, NATO, and ENISA.
- Anthropic is racing to establish safeguards as it expects rivals like OpenAI (with GPT-5.5-Cyber) to develop similar cybersecurity AI models soon.