Speaking to Information Week, SRI’s Ulf Lindqvist explained AI’s role in empowering cyber defenders to confront emerging threats.
When it comes to cybersecurity, AI is a double-edged sword. On the one hand, it’s actively being used by bad actors to penetrate sophisticated cyber defenses. On the other hand, cybersecurity experts recognize that it also provides them with new tools to counter bad actors. How can cybersecurity teams ensure that AI keeps them one step ahead?
Speaking to Information Week, SRI senior technical director Ulf Lindqvist pointed to some key areas where AI can aid cyber defenders, including in areas such as anomaly and malware detection.
“It’s probably necessary to use,” Lindqvist told Information Week, “because the attackers are using AI to boost their own productivity, to automate attacks, to make them happen and evolve faster than humans can react.”
As a senior technical director in SRI’s Computer Science Lab, Lindqvist manages R&D programs related to infrastructure security for both government and commercial clients and also leads SRI’s support for the U.S. Department of Homeland Security’s Science and Technology Directorate.
Given the cyber risks at stake in both industry and national defense, Lindqvist cautioned cybersecurity professionals against giving AI too much of a role in cases where the underlying systems to govern those AI capabilities are not yet mature. In particular, he pointed out that fully automated AI systems could make incorrect assumptions that render suspicious behavior invisible to an organization’s cybersecurity teams.
“What worries me is when you put AI in charge,” he said. “Putting the AI in charge of things when the technology is really not ready for that.”
SRI’s recent cybersecurity work includes research on malware embedded in large language models, high-assurance zero-knowledge cryptography, and cyberpsychology-informed network defense.
