New Bill Aims for AI Security at NSA
Representative Darin Lahood, a Republican from Illinois, has introduced a bill aimed at creating an “AI Security Playbook” at the National Security Agency (NSA). This initiative seeks to address the threats posed by China and other foreign challenges.
Dubbed the “Advanced AI Security Readiness Act,” the bill instructs the NSA’s artificial intelligence security center to develop strategies to tackle vulnerabilities in sensitive AI systems. It focuses on aspects like threat detection, cyber and physical security protocols, as well as contingency plans.
Lahood is collaborating with the House Select Committee on China, where Speaker John Mourenard plays a significant role. He mentioned this initiative marks the first time Congress has taken steps to formalize approaches to keep the U.S. ahead of the tech race against the Chinese Communist Party (CCP). The proposed legislation aims to enhance export controls, particularly on chip manufacturing, safeguard critical AI technologies with an emphasis on cybersecurity, and restrict investments in companies tied to the CCP or the People’s Liberation Army.
Lahood expressed concern about China’s ambitions, noting, “We start with the assumption that China has a plan to replace the U.S.” He stressed that AI is the “next frontier,” and although the U.S. is currently a leader in technology and AI, this bill would ensure proper safeguards are in place to protect national assets and encourage innovation.
He cited evidence from committee investigations indicating that the CCP has employed illegal distillation technology to illicitly extract insights from U.S. AI models, thereby accelerating its technological advancements. Additionally, he highlighted concerns about how AI chips could be smuggled to bypass U.S. export restrictions.
Lahood emphasized the importance of maintaining a competitive edge in AI, acknowledging that while experts currently believe the U.S. is ahead, this lead is not substantial. He believes the bill is crucial for protecting U.S. interests and ensuring that no AI-related exports could strengthen China’s capabilities.
“The future of cybersecurity and military capabilities hinges on who excels in the AI field,” he continued. “A win in the AI race will have far-reaching economic implications.” Lahood also mentioned the need to balance commercial interests with security, aiming to position U.S. tech firms strategically.
The bill outlines the NSA’s responsibility to identify vulnerabilities within AI data centers, developing methods to manage these risks effectively. This includes formulating strategies to thwart cyber threats specifically targeting AI technologies.
Although not directly connected, the legislative proposal comes on the heels of warnings from FBI Director Kash Patel regarding CCP operatives infiltrating U.S. agencies. Patel recently highlighted risks associated with Chinese nationals charged with smuggling potential biological weapons into the U.S.
Lahood stated that this case underscores the CCP’s extensive and sophisticated operations. He also mentioned the appropriation of U.S. farmland in strategic areas by the CCP, which raises significant national security concerns.
In his remarks, Lahood noted that while a fair competition would favor the U.S., China tends to operate under a different set of rules. “This isn’t just my perspective; there are several examples that highlight why legislation like this is crucial,” he said.
The bill is supported by the previous Trump administration, aiming to strengthen the U.S. AI infrastructure. Major tech companies, including Amazon and Nvidia, are making significant investments in AI data centers. For instance, Amazon recently unveiled a $20 billion investment in a rural Pennsylvania AI data center.
In late May, the NSA’s AI Security Center released guidance on “Risks and Best Practices for AI Data Security.” The recommendations focus on data protection methods, including using digital signatures for trusted revisions, tracking data origins, and leveraging secure infrastructures. This guidance is particularly pertinent for organizations that utilize AI systems in their operations, including the Department of Defense and other stakeholders within the defense industry.
