Senior officials in the Trump administration have recently grappled with how to regulate the rapidly evolving field of artificial intelligence, particularly focusing on the role of open-source AI models. A group including White House Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent, and Commerce Secretary Horward Lutnick has debated various potential interventionist measures targeting these models, which are freely available to download and modify. Discussions have included options such as imposing sanctions, placing Chinese AI companies on trade blacklists, or prohibiting U.S. cloud providers from engaging with such firms, according to multiple current and former government and tech insiders speaking on condition of anonymity.

However, amid pushback from Silicon Valley, the administration appears to have shifted toward an approach emphasizing support for American AI technologies rather than direct restrictions. In line with this approach, the White House recently invited U.S. technology firms to explore a framework under which the government could review new AI models for cybersecurity risks prior to public release. Details of this proposed review process remain limited.

The internal debate within the administration reflects the challenges faced by policymakers in responding to a swiftly developing technology with broad implications. Throughout President Trump’s second term, the administration has taken an active role in regulating various sectors, but the pace of AI innovation has complicated efforts to establish clear policy direction.

President Trump initially positioned himself as a proponent of AI, easing regulations and fostering close ties with industry leaders. That stance shifted in June when he issued an executive order aimed at increasing government oversight of AI models. The latest discussions have been fueled in part by the emergence of powerful new Chinese AI models distributed openly, revealing computer code or calculation methods, a practice known as “open weight.”

Within the U.S. technology sector, responses to Chinese open-source models have been mixed. Companies like OpenAI and Anthropic, which generally keep their underlying technologies private, have lobbied for stronger government actions against Chinese rivals, citing national security risks. Conversely, firms such as Nvidia, which produce open-source models or hardware for them, argue that openness fosters innovation and can enhance cybersecurity defenses.

The debate takes on heightened significance amid ongoing strategic competition between the United States and China over technological leadership. Chinese President Xi Jinping is expected to visit Washington in September, with AI likely to be a discussion topic. Industry analysts suggest that major U.S. policy moves on AI are unlikely before that visit.

Concerns have also centered on the practice of “distillation,” where new models are derived from the outputs of existing AI tools, potentially infringing on proprietary technology. The administration has publicly criticized Chinese companies for allegedly distilling U.S.-developed models, describing such actions as intellectual property theft.

Meanwhile, the administration is exploring regulatory mechanisms, including restrictions on the AI models that federal agencies can use and leveraging the Department of Commerce’s authority over information and communication technologies to limit U.S. business ties with certain Chinese AI firms.

Representatives of OpenAI have emphasized the importance of both open-weight and closed models in democratizing AI access and called for a coherent national framework to assess new models, manage associated risks, and equip U.S. cyber defenders with advanced tools.

Observers note that underlying much of the debate is a contest over market dynamics: whether AI development will be dominated by a few major players controlling proprietary systems or evolve into a more open ecosystem with broader participation. While national security concerns are present, many industry figures view the controversy primarily as a struggle over the future shape of the AI market.