Anthropic Launches Project Glasswing to Defend Against AI-Powered Cyber Threats
Anthropic announces Project Glasswing, a major coalition of tech and financial giants using advanced AI to find and fix critical software vulnerabilities before malicious actors can exploit them.
A coalition of major technology and financial companies forms Project Glasswing to secure critical software against advanced AI threats. Anthropic leads this initiative alongside Amazon Web Services, Apple, Google, Microsoft, NVIDIA, and several other industry giants. The project emerges as a direct response to the advanced capabilities of Anthropic's new Claude Mythos 2 Preview model.
The Claude Mythos 2 Preview model demonstrates an unprecedented ability to find and exploit software vulnerabilities, surpassing all but the most skilled human hackers. This AI system already discovers thousands of high-severity flaws across every major operating system and web browser. Anthropic warns that these dangerous capabilities will soon proliferate widely, creating severe risks for economies, public safety, and national security if left unchecked.
Project Glasswing aims to flip this dynamic by deploying the Mythos model strictly for defensive security purposes across the coalition. Anthropic provides up to $100 million in usage credits to launch partners and over 40 critical infrastructure organizations to scan and secure their systems. The company also donates $4 million directly to open-source security organizations as an urgent call for global collaboration in the fast-moving AI era.