The National Security Agency is standing up an artificial intelligence security center to maintain America’s advantages in that space by protecting the technology from cyber threats, the head of the agency said Thursday.

“The AI Security Center will become NSA’s focal point for leveraging foreign intelligence insights, contributing to the development of best practices, guidelines, principles, evaluation methodology, and risk frameworks for AI security. With an end goal of promoting the secure development, integration, and adoption of AI capabilities within our national security systems, and our defense industrial base,” Army Gen. Paul Nakasone said at a National Press Club event in Washington, D.C.

A recent NSA study on AI found the agency must focus more on securing it, Nakasone said, as “national security system owners and the defense industrial base are increasingly acquiring, developing, and integrating AI capabilities into defense systems, cybersecurity, and mission capabilities. Concurrently, adversaries are moving quickly to develop and apply their own AI, and we anticipate they will begin to explore and exploit vulnerabilities in US and allied AI systems.”

The center will be part of the NSA’s Cybersecurity Collaboration Center, bringing together all of the NSA’s AI security work under one effort, said Nakasone, who is also the commander of U.S. Cyber Command. The center will also work with companies to counter and prevent threats to their intellectual property.

Securing artificial intelligence entails “protecting AI systems from learning, doing, and revealing the wrong thing,” he said. “We must build a robust understanding of AI vulnerabilities, foreign intelligence threats to these AI systems, and ways to encounter the threat in order to have AI security. We must also ensure that malicious foreign actors can't steal America's innovative AI capabilities.”

Nakasone did not give a timeline on when the center will be fully operational.

The center is part of a series of NSA efforts to work more closely with other government agencies and private companies to stay updated on cyber threats—especially when it comes to elections.

Asked about AI—including deepfakes—influencing voting in the upcoming 2024 U.S. general election, Nakasone said people need to practice vigilance, and that his team is making sure they “understand the threat techniques of our adversaries”—which the center will help them do.

Introduction