AI Tools Could Enhance Cybersecurity if Properly Secured
AI technologies have the potential to significantly strengthen cybersecurity defenses, provided they are safeguarded against misuse, according to Richard Horne, the UK's top cyber official.
The emergence of AI models such as Claude Mythos has attracted global attention after its developer, Anthropic, revealed the tool's exceptional hacking capabilities.
Anthropic is currently limiting access to Mythos to assist governments, major technology firms, and financial institutions in securing their systems as the cybersecurity community prepares for the model's broader release.
Richard Horne, head of the National Cyber Security Centre (NCSC), stated that advanced AI tools could represent a "net positive" for public cybersecurity if the technology is protected from exploitation.
This statement coincides with calls from the UK's Security Minister for AI companies to collaborate with the government on enhancing national cyber-defense capabilities.
Anthropic, known for its chatbot Claude, has not announced a release date for Mythos. However, the company raised widespread concern by claiming Mythos is an expert hacker, performing as well as or better than top human hackers.
The primary concern is that if Mythos falls into malicious hands or operates autonomously in a harmful manner, it could trigger significant data breaches or crippling cyberattacks.
Positive Outlook on AI at CyberUK Conference
During a speech at the NCSC's annual CyberUK conference on Wednesday, Horne presented a more optimistic perspective, arguing that AI tools can contribute to safer and more secure cyber environments.
He encouraged organizations and companies not to fear emerging AI-driven attacks but to ensure fundamental cybersecurity practices are rigorously implemented.
"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," he said.
Horne's remarks reinforce previous advisories emphasizing the urgency for individuals and organizations to update software and modernize legacy IT systems.
He also urged AI developers to secure their models by adhering to newly established European safety guidelines.
Government Calls for AI Industry Collaboration
At the same event, Security Minister Dan Jarvis appealed to AI companies to partner with the government in a "generational endeavour" aimed at ensuring AI technologies are leveraged to protect critical infrastructure from cyber threats.
Currently, all leading frontier AI models are developed outside the UK, predominantly by companies based in the United States and China.
This situation means the UK depends on firms like Anthropic for access to tools such as Mythos and lacks direct control over their development, training, or deployment.
OpenAI has also introduced a cybersecurity-focused AI model named GPT 5.4 Cyber, which it claims possesses strong capabilities.
Ongoing Cyber Threats and National Security
The CyberUK conference speeches also highlighted the persistent threat posed by nation-state and hacktivist cyberattacks, particularly those originating from Russia and China.
The NCSC emphasized that cyber operations have become the "home front" of defense in the UK, with recent incidents such as attacks linked to Iran demonstrating the increasing significance of cyber warfare in modern conflicts.
for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? here.
