In a significant move to protect national security infrastructure, US lawmakers have introduced legislation designed to create a digital firewall against potentially hostile artificial intelligence systems.
The bipartisan 'No Adversarial AI Act,' introduced on June 25, 2025, would prohibit federal agencies from using AI models developed in countries considered adversaries of the United States, including China, Russia, Iran, and North Korea. The bill specifically targets Chinese AI systems like DeepSeek, which has raised concerns after a senior US official concluded it was aiding China's military and intelligence operations.
"The U.S. must draw a hard line: hostile AI systems have no business operating inside our government," said Representative John Moolenaar (R-MI), who chairs the Select Committee on the Chinese Communist Party. "This legislation creates a permanent firewall to keep adversary AI out of our most sensitive networks - where the cost of compromise is simply too high."
The bill has garnered support across party lines, with co-sponsors including Representatives Raja Krishnamoorthi (D-IL), Ritchie Torres (D-NY), and Darin LaHood (R-IL) in the House, while Senators Rick Scott (R-FL) and Gary Peters (D-MI) are leading the companion measure in the Senate.
Under the proposed legislation, the Federal Acquisition Security Council would be required to develop and regularly update a public list of AI models developed by foreign adversaries. Federal agencies would be prohibited from acquiring or using these AI systems, with limited exceptions for research, counterterrorism, or mission-critical functions.
This regulatory move represents a significant expansion of tech decoupling policies, extending previous restrictions on hardware companies like Huawei to now include AI software and algorithms. The legislation comes amid growing concerns about AI as a potential vector for espionage, data theft, and critical infrastructure sabotage, marking an important shift in how Western countries are approaching AI security governance.