Legislation introduced Wednesday in Congress aims to ban all Chinese artificial intelligence systems from federal government use, citing national security threats.
The bipartisan proposal, pushed by both House and Senate members, would bar any AI developed by U.S. adversaries, notably China, from operating in federal agencies.
The bill makes rare exceptions for counterterrorism and research purposes, giving the government a free pass to use the tech when convenient.
The legislative panic follows the release of DeepSeek, a Chinese AI model nearly matching OpenAI and Google in power—despite being built on a budget.
Lawmakers say the U.S. must win what they describe as a “new Cold War,” with AI now framed as the world’s most important battleground.
The proposal coincides with increasing concern over China’s growing AI capabilities, as Stanford’s 2025 AI Index shows China catching up in performance and leading in publications and patents.
Lawmakers argue that authoritarian-built AI could “infect” democratic institutions, assuming that technology inherits political values like an ideological virus.
Committee hearings featured calls for stricter export controls on advanced chips, with experts claiming that without limits, China could use U.S. technology to “build powerful AI to harm American interests.”
The bill would empower the Federal Acquisition Security Council to identify AI systems considered high-risk, although no one seems to know how many are already in use.
Federal agencies and some states have reportedly banned DeepSeek from government devices, but the new bill seeks to enforce a sweeping, standardized prohibition.
Despite the sense of urgency, experts note loopholes in existing chip export controls that continue to allow China access to restricted technologies.
Lawmakers insist that AI developed by authoritarian regimes has “no business” operating in American government, even though the technology has likely already slipped through the cracks.
As the AI arms race intensifies, Congress appears committed to securing America’s digital turf—just as soon as they finish figuring out what AI actually is.