Discord users breach access controls to reach Anthropic’s Mythos model

A small group of Discord users allegedly bypassed controls and accessed Anthropic’s restricted Mythos model through a third-party environment.

The model had only recently been released in a tightly limited form, and there is no confirmed evidence of malicious use.

The episode highlights how AI security depends not just on the model, but also on vendors, permissions, and surrounding infrastructure.

It also raises concerns for dual-use cybersecurity tools, where unauthorized access could increase risk even without clear misuse.