A group of unauthorized users has reportedly gained access to Mythos, Anthropic's new cybersecurity tool. The tool was designed as an enterprise security solution but now faces the risk of misuse.

Anthropic confirmed the report, stating they are investigating and so far have found no evidence that their systems were impacted. However, the group has been using the software since its public announcement and provided evidence to Bloomberg in the form of screenshots and a live demonstration.

Mythos was released to select vendors as part of Project Glasswing, aiming to prevent bad actors from using it. But if true, this unauthorized use could spell trouble for Anthropic's security assurances.

The incident raises questions about how secure AI tools can be in the wrong hands and whether they can truly protect against cyber threats. It highlights the potential risks inherent in sharing cutting-edge technology with third parties.