Anthropic Launches ‘Project Glasswing’ with Big Tech to Advance AI Cybersecurity

Anthropic has unveiled a new cybersecurity initiative, Project Glasswing, partnering with major technology firms including Amazon, Microsoft, and Apple to test advanced AI-driven security capabilities.

As part of the initiative, selected organizations will gain early access to Anthropic’s unreleased model, Claude Mythos Preview, designed to support defensive cybersecurity efforts. Additional partners include CrowdStrike, Palo Alto Networks, Google, and Nvidia.

Anthropic said the model has already identified thousands of vulnerabilities across operating systems, web browsers, and other software, highlighting its potential to strengthen cyber defense strategies. Participating organizations will use the tool to enhance their security operations, while Anthropic plans to share insights with the broader industry.

The announcement comes amid growing concerns over AI-powered cyber threats, a key topic at this year’s RSA Conference in San Francisco. Recent reports have also noted that advanced AI models like Claude Mythos could introduce new risks even as they offer powerful capabilities.

To expand the initiative, Anthropic is extending access to around 40 additional organizations responsible for critical software infrastructure. The company has also committed up to $100 million in usage credits and $4 million in funding to support open-source security efforts.

Anthropic said it is working closely with the U.S. government to evaluate the model’s capabilities and ensure safe deployment. The company’s long-term goal is to enable organizations to securely scale the use of advanced AI systems like Mythos.

The initiative follows earlier incidents where vulnerabilities in Anthropic’s Claude AI were exploited in attacks on multiple global organizations. Industry data also underscores the urgency: a study by IBM and Palo Alto Networks found that 67% of executives reported being targeted by AI-driven cyberattacks within the past year.

With Project Glasswing, Anthropic is positioning AI not just as a source of new risks, but as a critical tool in defending against the next generation of cyber threats.