Anthropic launches Project Glasswing to test advanced AI for cybersecurity
Indian Express·3h ago·Positivity
What Happened
Anthropic introduced Project Glasswing, a cybersecurity initiative with major tech firms and infrastructure organisations testing an unreleased AI model to identify software vulnerabilities. The system, called Claude Mythos Preview, has reportedly found serious flaws in operating systems and browsers that went undetected for years. Access remains restricted to closed collaborators. Microsoft, AWS, and Google participate. Anthropic committed $104 million toward security efforts.
Key Entities
AnthropicProject GlasswingClaude Mythos PreviewMicrosoftAmazon Web ServicesGooglecybersecurityAI vulnerabilities