Treasury Secretary Scott Bessent has told lawmakers that the United States must foster open artificial intelligence (AI) models to remain competitive against Chinese firms. During his testimony before the House Financial Services Committee on Tuesday, Bessent addressed the economy and international finance, with artificial intelligence emerging as a central topic of discussion.
When questioned about advancements in AI and the progress toward artificial general intelligence, Bessent highlighted Anthropic’s Mythos model, released earlier this year, as a significant step forward. He emphasized that the U.S. needs to develop more open-source AI models to drive innovation and prevent regulatory capture by large labs.
“One of the things I would say is that we need to develop… more open-source models in the U.S. We can’t let these large labs have regulatory capture because that will stop innovation,” Bessent told the panel.
Secretary Scott Bessent talks to Bret Baier, Fox News anchor, during an Economic Club of New York event in New York, on June 23, 2026. (Krisanne Johnson/Bloomberg via Getty Images)
Open-source AI models allow the public to inspect codebases, review training parameters, or utilize them for various projects. In contrast, closed-source models are controlled by their developers and require specific access. AI labs in China have utilized a process known as distillation to train their open-source models on U.S. tech companies’ closed-source models, enabling them to build advanced models at a reduced cost. For instance, an open model from a Chinese developer was previously used in a breach where an OpenAI model escaped containment and compromised Hugging Face.
Open-source AI models can be trained through the illegal distillation of closed AI models. (Jaap Arriens/NurPhoto via Getty Images)
Bessent explained that developing domestic open-source models is a strategic way to counter Chinese advancements. He noted that Chinese models often distill U.S. models—which he described as a polite scientific term for “steal.” By developing domestic open models, the appeal of Chinese alternatives will diminish.
“One of the ways we can push back against China, because the Chinese models, they distill from the U.S. models – which is a polite, scientific word for steal – and the more we develop our own open source models here, or open models here, then… eventually the Chinese models, people will not use them,” Bessent stated.
He added that due to distillation, many Chinese models have come to believe they are on par with advanced U.S. models like Mythos or Claude. “Many of the Chinese models think they’re Mythos, they think they’re Claude,” Bessent remarked.
Major U.S. tech companies have opposed government restrictions on open-weight AI models, arguing that open models accelerate innovation, bolster cybersecurity, and ensure the American AI industry’s competitiveness. In July, a coalition including Nvidia, Microsoft, Meta, Dell Technologies, Palantir, Hugging Face, Mozilla, and Mistral signed a joint letter arguing against stifling regulations.
The Trump administration is working with AI companies and leaders like OpenAI CEO Sam Altman as they develop regulations for the emerging industry. (Ludovic Marin/Getty Images)
The joint letter stated that while unlawful extraction of value from closed models raises legitimate concerns, these issues should be resolved through targeted legal and commercial frameworks rather than broad restrictions on techniques vital to AI innovation.
The Trump administration’s voluntary AI framework, released in August, exempted open-source and open-weight models from mandatory pre-release security reviews, focusing instead on proprietary closed models.
