Anthropic launched Project Glasswing to proactively identify and repair critical software vulnerabilities. The cybersecurity initiative includes partners Amazon Web Services, Apple, Google, Microsoft, and NVIDIA. This coalition aims to defend against the rising threat of AI-powered cyberattacks.
The project utilizes Anthropic’s unreleased Claude Mythos Preview model. This model possesses advanced capabilities to discover and exploit software flaws. Project Glasswing prioritizes defensive applications of this technology to prevent dangerous proliferation.
Anthropic is committing up to $100 million in model access credits to partners and software organizations. The company is also donating $4 million to support open-source security efforts. Markets were closed at the time of the Sunday announcement, resulting in no immediate price reaction.