Anthropic Says Its New AI Model Is So Good at Finding Security Risks, You Can’t Use It

Source Domain: www.cnet.com

Anthropic’s newest AI model, Claude Mythos, is exceptionally good at identifying cybersecurity vulnerabilities, posing a substantial threat to public release.
The technology has already discovered thousands of severe security vulnerabilities in major operating systems and web browsers.
To mitigate risks, Anthropic is collaborating with major tech companies through Project Glasswing to utilize the model for patching vulnerabilities.
The consortium includes notable organizations like Apple, Amazon Web Services, Google, and Microsoft, among others, and aims to safeguard AI-powered cybersecurity.
Anthropic is providing $100 million in usage credits for Mythos and $4 million in donations to open-source security initiatives.
The initiative involves over 40 organizations collectively working to strengthen defenses against AI-based cyber threats.
This development underscores the broader implications and challenges of AI in cybersecurity, prompting national policymakers to consider regulatory measures.
The project could serve as a model for addressing the balance between advancing AI technology and ensuring public safety.

You may have missed