Claude Mythos AI unauthorised access claim probed by Anthropic
bbc.com
April 22, 2026
3 min read
🔥🔥🔥🔥🔥
29/100
Summary
Anthropic is investigating a claim of unauthorized access to its Claude Mythos AI tool, which is described as too powerful for public release. The investigation was prompted by a report indicating that a small group accessed the tool through a third-party vendor environment.
Key Takeaways
- Anthropic is investigating unauthorized access claims related to its Claude Mythos AI tool, which is considered too powerful for public release.
- Access to the Mythos model was reportedly gained through misuse of permissions rather than a traditional hacking incident.
- The UK's National Cyber Security Centre emphasizes the potential of advanced AI tools to improve security, while also warning about the risks of unauthorized access leading to malicious activities.
- The UK relies on foreign companies like Anthropic for access to advanced AI models, raising concerns about control over their development and deployment.
Related Articles
'Too Dangerous to Release' Is Becoming AI's New Normal
Apr 25, 2026
US summons bank bosses over cyber risks from Anthropic's latest AI model
Apr 10, 2026
A leak reveals that Anthropic is testing a more capable AI model "Claude Mythos"
Mar 27, 2026
After dissing Anthropic for limiting Mythos, OpenAI restricts access to Cyber
May 1, 2026
Opus 4.6 uncovers 500 zero-day flaws in open-source code
Feb 5, 2026