U.S. Bipartisan Bill Seeks to Ban "Adversarial AI" from Foreign AdversariesU.S. bipartisan lawmakers introduced a new bill, the No Adversarial AI Act, on June 25 local time. It aims to restrict federal government procurement and use of AI technologies developed by "foreign adversaries" like China, strengthening technical security for national critical systems.

The bill’s core provisions include creating a "Federal Adversarial AI List" and banning federal agencies from using AI products on the list, with limited exemptions for research, security, and other specific purposes. It explicitly designates China, Russia, and Iran as "foreign adversaries"—aligning with recent calls by U.S. lawmakers to fully ban Chinese firm DeepSeek’s models. This reflects an escalation of U.S. strategy to contain China in AI and promote technological decoupling.

If passed, the bill could further intensify camp-based competition in the global AI sector, posing severe challenges to the research environment. While currently limited to federal agencies, expansion to the commercial sector cannot be ruled out in the future.

The move underscores growing U.S. vigilance over AI supply chain risks. However, critics warn it may fragment global AI collaboration, hinder technological progress, and force companies into costly adjustments to avoid regulatory barriers.