TLDRs:
- A bipartisan bill aims to ban federal use of AI from China, Russia, Iran, and North Korea.
- DeepSeek, a Chinese AI firm, is at the center of U.S. national security concerns.
- The proposed law would allow AI models to be removed from the ban list if proven free of foreign control.
- Lawmakers say the move institutionalizes the ongoing tech decoupling from geopolitical rivals.
In a bold move to tighten the U.S. government’s digital defenses, a bipartisan group of lawmakers has introduced a new bill seeking to ban artificial intelligence tools developed in adversarial nations from federal use.
The “No Adversarial AI Act,” unveiled on Wednesday, underscores growing unease over AI technologies originating in countries like China, Russia, Iran, and North Korea. While tensions around semiconductor exports have dominated headlines in recent years, this bill shifts the focus to AI software, highlighting how Washington is expanding its tech containment strategy beyond hardware.
The legislation arrives amid heightened scrutiny of DeepSeek, a Chinese AI company previously linked to military and intelligence institutions. According to a report from the House Select Committee on China released in April, DeepSeek’s technology allegedly transmits U.S. user data back to China and modifies outputs to align with Chinese censorship norms.
The firm’s ties to state-backed organizations and its founder’s ecosystem have raised red flags within Congress, catalyzing a coordinated legislative response.
Bipartisan unity across both chambers
The bill was introduced simultaneously in both the House and Senate, with lawmakers from both parties expressing concern over the risk of foreign-controlled AI infiltrating critical government infrastructure. Representatives John Moolenaar and Raja Krishnamoorthi spearheaded the House version, while Senators Rick Scott and Gary Peters are leading efforts in the Senate. The initiative has also garnered backing from other lawmakers, including Ritchie Torres and Darin LaHood, reflecting rare bipartisan consensus on an emerging technology issue.
Representative Moolenaar emphasized that the bill is designed to protect the U.S. government from “hostile AI systems that could compromise national security.” Under the proposal, federal agencies would be prohibited from deploying AI models linked to countries deemed adversarial unless cleared by Congress or the Office of Management and Budget. However, the bill includes a provision for AI systems to be removed from the restricted list if evidence shows they are no longer under foreign influence.
A broader shift in tech policy
The proposed act represents more than just a reaction to DeepSeek or a handful of AI companies. It marks a broader transformation in U.S. policy from managing tech risks to proactively shaping a secure digital future. In recent years, Washington has imposed increasingly stringent export restrictions on advanced chips, most notably targeting China’s access to NVIDIA’s high-performance processors.
Now, with AI models themselves becoming a national security focus, lawmakers are extending the perimeter of protection into software.
What sets this bill apart is its attempt to create a permanent legal framework for tech decoupling, rather than relying on temporary executive orders or case-by-case sanctions. By embedding restrictions in federal law, Congress aims to institutionalize its approach to safeguarding critical systems against foreign AI infiltration.
China’s AI gains drive urgency
The move also comes as Chinese AI models make rapid strides, with performance benchmarks now trailing top U.S. models by only a few months. Despite U.S. export controls, China’s AI sector has shown resilience, with advances fueled by state investment and a tightly coordinated push across government, academia, and industry.
That growing competitiveness has sharpened U.S. efforts to draw firmer boundaries around the use of potentiall compromised technologies.
As geopolitical and technological rivalries continue to converge, the No Adversarial AI Act may prove to be a landmark moment in how the United States defines and defends its digital sovereignty.
