Claude Mythos will not be available to the general public soon. Anthropic, the creator of Claude Mythos, an AI model adept at identifying security flaws in software, has decided against public release due to potential misuse concerns. The details about Mythos were leaked due to a security issue in Anthropic’s systems. Although Claude Mythos Preview, a promising AI model for cybersecurity, was announced with potential to find high-severity vulnerabilities, the risks of misuse led Anthropic to limit access to select organizations. This initiative, called “Project Glasswing,” aims to enhance critical software infrastructure. The company’s decision followed internal deliberations, focusing on providing cybersecurity defenders with strategic advantages. Anthropic CEO Dario Amodei emphasized controlled access to address vulnerabilities before such AI models become widespread. While Claude Mythos is proficient in cybersecurity, its broader release remains uncertain as the company explores safe deployment strategies.
Anthropic Restricts Access to Claude Mythos AI Identifying Security Flaws
1 Min Read