The UK's top cybersecurity official has declared that advanced artificial intelligence tools like Claude Mythos could ultimately benefit public safety despite their hacking capabilities, provided they are properly secured.
Richard Horne, head of the National Cyber Security Centre (NCSC), will argue in a speech at the CyberUK conference that frontier AI models present a "net positive" opportunity for cyber defenses. This comes as Anthropic, the creator of the popular Claude chatbot, revealed its new Mythos model demonstrates expert-level hacking skills comparable to top human specialists.
"As we have seen in the media in recent days, frontier AI is rapidly enabling discovery and exploitation of existing vulnerabilities at scale, illustrating how quickly it will expose where fundamentals of cyber-security are still to be addressed," Horne will tell conference attendees.
Anthropic has restricted access to Mythos while collaborating with governments, technology companies, and financial institutions to fortify systems ahead of its broader release. The company has not specified a timeline for public availability.
Security Minister Dan Jarvis will also address the conference, urging AI developers to partner with the government on what he calls a "generational endeavor" to protect critical infrastructure. Jarvis emphasized the need for international cooperation, noting that all leading frontier AI models originate outside the UK, primarily from US and Chinese firms.
The NCSC warns that cyber threats now represent "the home front" of national defense, with recent incidents like attacks linked to Iran demonstrating how digital conflicts increasingly shape modern warfare. Officials highlighted ongoing risks from state-sponsored actors and hacktivist groups, particularly those affiliated with Russia and China.
Horne's message focuses on practical preparedness rather than alarm, encouraging organizations to strengthen basic cybersecurity practices like regular software updates and legacy system modernization. He also advocates for AI companies to adhere to newly established European safety standards to prevent misuse of their technologies.
While acknowledging legitimate concerns about powerful AI tools falling into malicious hands, UK officials maintain that properly managed AI could significantly enhance the nation's ability to identify and patch vulnerabilities before they can be exploited by adversaries.