Anthropic Launches AI Cybersecurity Project With Tech Giants

Anthropic has unveiled Project Glasswing, an initiative giving major tech partners including Amazon, Microsoft, Apple, Google, and Nvidia early access to its unreleased AI model Claude Mythos Preview for defensive cybersecurity work. The announcement follows reports that the model posed security risks alongside advanced capabilities, which had previously rattled cybersecurity stocks. Anthropic says Mythos Preview has already identified thousands of major vulnerabilities in operating systems and browsers. The company is extending access to 40 additional organizations, committing up to $100 million in usage credits and $4 million in donations to open-source security groups, while holding ongoing discussions with the US government about the model's capabilities.