Anthropic is on high alert as reports suggest unauthorised users may have accessed its powerful Claude Mythos model. The tool was meant for select tech and financial firms but now faces scrutiny over security.
The incident highlights the challenges large AI companies face in preventing advanced tools from falling into dangerous hands. While no malicious use has been detected yet, experts warn of potential misuse, such as fraud or cyber abuse.
At a recent CyberUK conference, UK's National Cyber Security Centre urged delegates to focus on basic security measures. However, the UK's reliance on foreign AI firms for cutting-edge technology underscores the ongoing vulnerability.
The event also highlighted the increasing threat from nation-state attacks, with Russia and China singled out as particular concerns. The NCSC stresses that cyber defence is now a crucial aspect of modern conflict.
