Treasury Secretary Bessent Advocates for Open-source AI in US
WHO:
Techmeme - Treasury Secretary Scott Bessent testified before the House Financial Services Committee.
Fox Business - Bessent discussed the need for open-source AI models to compete with Chinese firms.
WHAT:
Techmeme - Bessent stated that AI labs should not receive liability exemptions.
Fox Business - Bessent emphasized the need for more open-source AI models in the U.S. to foster innovation and competition.
Fox Business - Bessent noted that Anthropic's Mythos model, released earlier in the year, represented a significant advancement in AI.
Fox Business - Bessent quoted, "One of the things I would say is that we need to develop… more open-source models in the U.S."
Fox Business - Bessent remarked that Chinese models are distilling from U.S. models, which he described as "a polite, scientific word for steal."
WHERE:
Techmeme - The testimony took place at a U.S. House hearing.
Fox Business - The discussion occurred in Congress.
WHY:
Fox Business - Bessent argued that developing open-source models in the U.S. is essential to counteract the advancements made by Chinese AI firms.
Fox Business - Bessent stated that the lack of regulatory capture by large labs is crucial for innovation.
HOW:
Fox Business - Bessent explained that open-source AI models allow public inspection of codebases and training parameters, contrasting with closed-source models controlled by developers.
Fox Business - Bessent highlighted that Chinese firms have trained open-source models using U.S. tech companies' closed-source models through a process known as distillation.
(UPDATED September 16, 2026 01:59 UTC)
Treasury Secretary Scott Bessent testified before the House Financial Services Committee, emphasizing the need for more open-source AI models in the U.S. to foster innovation and compete with Chinese firms. He argued that AI labs should not receive liability exemptions and noted that Chinese models are essentially distilling from U.S. models, which he described as "a polite, scientific word for steal." Bessent highlighted the significance of open-source models, explaining that they allow for public inspection of codebases and training parameters, contrasting them with the closed-source models controlled by developers.