AI's role in uncovering software vulnerabilities raises cybersecurity concerns
Researchers at the University of Groningen warn that advanced AI systems are accelerating the discovery of software vulnerabilities, intensifying cybersecurity challenges for organizations.
Experts at the University of Groningen have voiced concerns that advanced artificial intelligence systems are intensifying the cybersecurity "cat and mouse game" between organizations and malicious actors.
Aline Klingenberg and Evgeni Moyakine, researchers specializing in cybersecurity and IT law, stated that AI can expedite the identification of software vulnerabilities. Their comments follow claims by Anthropic regarding its new generative language model, Mythos, which reportedly can expose weaknesses in operating systems, browsers, and servers.
Moyakine noted that while such AI tools can aid companies in detecting vulnerabilities and complying with regulations like the EU's NIS2 Directive, they also present risks. He cautioned that the same capabilities could be misused by hackers, posing new challenges to digital security.
The researchers also highlighted differing approaches to technology regulation between the US and Europe. They observed that Europe tends to assess new technologies earlier in terms of privacy, security, and societal impact, contrasting with the US emphasis on entrepreneurial freedom. However, they acknowledged the difficulty of enforcing digital regulations globally.
Moyakine and Klingenberg urged organizations to adapt their business processes and patch management strategies to stay ahead of AI-enabled threats. Controlling access to and the potential misuse of powerful AI tools like Mythos are seen as critical issues for future cybersecurity.