After identifying vulnerabilities across ALL major software platforms, Anthropic says Mythos will be tested in partnership with several industry partners.

Anthropic said in a statement that its new AI technology, Claude Mythos2 Preview, or “Mythos,” was recently tested to identify security vulnerabilities. It worked so well that it found thousands of cyber vulnerabilities, including those in “every major operating system and web browser.” Some of the exposed vulnerabilities go back over 25 years. 

According to Anthropic, the findings show that AI systems can significantly accelerate the process of identifying exploitable flaws, including those that could be used in cyberattacks. The company stated that making Mythos broadly accessible could enable cybercriminals and other bad actors to find and exploit vulnerabilities at scale.

Project Glasswing

Instead, Anthropic is giving Mythos access to some of the tech industry’s largest companies to “find and fix vulnerabilities or weaknesses in their foundational systems – systems that represent a very large portion of the world’s shared cyberattack surface.” This partnership, called Project Glasswing, includes the top five most valuable companies –  NVIDIA, Apple, Google, Microsoft, and Amazon, as well as cybersecurity giants such as CrowdStrike and Palo Alto Networks.

The company will also give access to an additional 40 companies that “build or maintain critical software infrastructure so they can use the model to scan and secure first-party and open-source systems.” 

Further, Anthropic committed $100M in usage credits earmarked for Mythos testing and $4M in donations to open-source security organizations.

The company expressed hope that Glasswing marks the beginning of large-scale collaborations with more private and public-sector entities to address AI’s impact on security.