Congressman Bill Huizenga has introduced legislation to help combat the theft of American AI technology by foreign countries.

Huizenga tells us the bill seeks to address attacks in which companies or governments copy AI components like model weights and architectures to cheaply build competing systems.

“It’s called model extraction, or distillation,” Huizenga said. “This has kind of become the new frontier of Chinese economic coercion and theft of U.S. intellectual property. And so we’re trying to push back at that.”

Firms like OpenAI, Google, and Anthropic have reported facing that threat. Huizenga says foreign actors, in particular China, are known to engage in the practice.

“Whether it’s traditional manufacturing or now this high-tech manufacturing and this high-tech intellectual property, China is best at going in and basically stealing it. That’s how they get around patents. That’s how they get around a lot of legal requirements about what can or can’t get shipped into their country.”

Huizenga’s “Deterring American AI Model Theft Act” would establish a framework to identify foreign actors extracting American models, allow discretionary sanctions, and create a name and shame list of AI thieves. He says it would also create a pathway for the U.S. State Department to work with private industry to share best practices and study attacks to strengthen defense.

Huizenga’s bill is now before the House Committee on Foreign Affairs.