Anthropic’s Mythos Model Is Being Accessed by Unauthorized Users
Anthropic is investigating reports that a small group of unauthorized users in a private online forum accessed and used its Mythos AI model, which is designed to identify and exploit software vulnerabilities. The access reportedly came through a mix of methods, including credentials from a third-party contractor and online sleuthing tools. The users were part of a private Discord channel that focuses on hunting for data about unreleased models.