Anthropic launches new AI model for cybersecurity

Anthropic has introduced Project Glasswing, a new AI model named Claude Mythos Preview designed for cybersecurity. This model aims to identify system vulnerabilities with minimal human intervention, partnering with major tech companies like Nvidia, Google, Amazon, Apple, and Microsoft. Due to its advanced capabilities and potential security risks, Anthropic is not planning a public release of Mythos, offering it only to select partners for defensive security work. The initiative seeks to proactively patch critical software and secure infrastructure before malicious actors can exploit flaws.