Archive — US — techcrunch
Unauthorized group has gained access to Anthropic’s exclusive cyber tool Mythos,report claims
Unauthorized users have reportedly accessed Anthropic’s AI cybersecurity tool, Mythos, via a third-party vendor. While the group has demonstrated the tool's capabilities, Anthropic is investigating the breach but claims there's no evidence of impact on their systems. Mythos was initially released to control its usage and to ensure enterprise security.
Key points
- Unauthorized group accessed Anthropic's Mythos tool through a third-party vendor.
- The access was gained on the same day Mythos was publicly announced.
- Members are part of a Discord channel focused on unreleased AI models.
- The group provided screenshots and a live demo to Bloomberg as proof of access.
- Anthropic is investigating the breach but has found no adverse impact on its systems so far.
- Mythos was designed for enterprise security but could be weaponized if misused.
Why it matters
- The incident reflects rising concerns about cybersecurity around powerful AI tools.
- Increasing incidents of unauthorized access to sensitive technologies mirror broader cybersecurity challenges reported in 2025.
- Previous cases of tool misuse (e.g., exploits of AI models) have led to stricter release protocols by tech companies, similar to Anthropic's initial approach.
The reported unauthorized access to Mythos raises significant concerns regarding the security of AI tools and emphasizes the need for stringent measures to prevent misuse in a rapidly evolving tech landscape.