Chinese AI Models Spark Divisions Among Trump's Advisors
Recent launches of free, intelligent Chinese AI models, notably Moonshot's Kimi, have intensified public disputes among current and former advisors to President Donald Trump on artificial intelligence. These models are seen as posing economic and political challenges to the administration by offering alternatives to costly US-developed AI. The situation has led to public criticisms of leading American AI companies and highlighted differing views within Trump's circle regarding government intervention, open-source competition, and national security implications of rapidly evolving AI technology.
Public disagreements have emerged among current and former artificial intelligence advisors to President Donald Trump regarding the impact of Chinese AI models on the US market and national security.
Over the weekend, David Sacks, who served as Trump's AI and crypto “czar” until March, publicly criticized Anthropic's models as “lobotomized” and “woke.” Emil Michael, a top Pentagon official, referred to OpenAI’s new head of strategic futures as a “supreme village idiot.” These comments surfaced following the launch of Kimi, a free, open-source model developed by Chinese AI company Moonshot, which reportedly rivals the intelligence of models from US companies like OpenAI and Anthropic.
The emergence of Kimi and similar Chinese models presents both economic and political challenges for the Trump administration. US companies may find less incentive to invest in models from Anthropic or OpenAI when free alternatives are available. This trend, according to Anton Leicht, a fellow at the Carnegie Endowment, could be a "threat for an administration that really doesn’t want more economic bad news," and has already impacted US stocks.
David Sacks, no longer formally advising Trump, has argued against government intervention to protect leading AI companies from open-source competition. He suggested that Chinese AI models gain popularity due to fewer restrictions on usage, despite inherent state censorship. His position contrasts with a prevailing view within the administration that emphasizes a larger role for government intervention due to potential national security threats posed by powerful AI models.
This latter perspective has fueled a new White House review process designed to vet the security of AI models before their release. Dean Ball, a former Trump AI advisor now with OpenAI, criticized this initiative as a “de facto licensing regime for frontier AI.” Ball speculated that the administration might use subtle pressure to discourage US companies from utilizing models like Kimi. Michael, however, strongly opposed this idea, advocating for a democratic process over what he termed a “Deep State scheme.”
The discussion also touches upon how Chinese models achieve such advanced capabilities. Despite efforts to restrict China's access to top chips during the Biden administration and early in Trump's second term, export controls have eased. Trump reportedly allowed Nvidia to sell more chips to China in exchange for a US government cut, though some chip smuggling has also been alleged. China's computing power is noted as limited. According to MIT Technology Review, this situation complicates efforts to address the rise of advanced Chinese AI.
