technology
Anthropic Restricts Claude Mythos AI Access to Select Organizations After Hacking Demonstration
Anthropic has revealed its AI model Claude Mythos excels at hacking and is restricting access to governments, tech giants, and banks. UK cyber officials, including NCSC head Richard Horne, stated that such advanced AI can enhance public cyber-security if secured from misuse. Security Minister Dan Jarvis urged AI firms to collaborate with the government on national cyber-defense.
Substrate placeholder — needs reviewAudio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
Anthropic revealed that its AI model Claude Mythos is extremely good at hacking. The company claimed that Claude Mythos is an expert hacker as good as, if not better than, the best humans. Anthropic is restricting access to the Claude Mythos model to governments, tech giants, and banks.
Anthropic has not announced a release date for its newest model Mythos. The company, which is the maker of the chatbot Claude, sparked these developments as the cyber-security world prepares for potential impacts. BBC News reported that the threat of AI such as Claude Mythos has made headlines around the world.
Richard Horne, head of the National Cyber Security Centre (NCSC), stated that advanced AI tools can be a 'net positive' to public cyber-security if secured from misuse. Horne will speak at the NCSC's annual conference CyberUK on Wednesday. In his speech, Horne will argue that AI tools can make things safer and more secure.
Horne will urge companies and organisations to ensure they are doing the basics of cyber-security right. ' Horne urged people to update software on their systems and upgrade legacy IT. Horne also urged AI companies to follow newly-created European safety guidelines for their models.
The UK's security minister urged AI companies to work with the government on national cyber-defence capabilities. Security Minister Dan Jarvis will speak at the CyberUK event. Jarvis will implore AI firms to work with the government on using AI to protect critical networks from attackers.
