isnaini azizah
The recent public spat began with prominent Trump advisors launching scathing critiques of leading U.S. AI companies. David Sacks, who previously served as Trump’s "AI and crypto czar," decried Anthropic’s models as "lobotomized" and "woke." Simultaneously, Emil Michael, a senior Pentagon official, labeled OpenAI’s new head of strategic futures a "supreme village idiot." These vitriolic exchanges are not merely personal animosities; they are symptomatic of a broader strategic disagreement triggered by the rapid advancements and accessibility of Chinese AI technology.
At the heart of the turmoil is Kimi, a free, open-source AI model from Chinese company Moonshot, which has demonstrated capabilities rivaling those of expensive, proprietary models from OpenAI and Anthropic. The existence of such powerful and accessible alternatives poses a direct threat to the business models of U.S. AI firms, which have become significant drivers of economic growth. As Anton Leicht, a fellow at the Carnegie Endowment, observed, "a threat for an administration that really doesn’t want more economic bad news." The market has already reacted, with Chinese AI models reportedly rattling U.S. stocks.
This situation presents a complex conundrum for Donald Trump, especially as it unfolds against a backdrop of increasing public distrust towards AI companies and growing concerns about the energy demands of AI infrastructure, exemplified by New York’s recent ban on new data centers. A significant portion of the American public, likely to be unsympathetic to the financial struggles of large tech companies, may not view government intervention to protect these firms as a priority.
Interestingly, David Sacks, despite his harsh words for U.S. AI companies, has voiced a sentiment that might resonate with some of this public skepticism. He criticized top AI companies for "want[ing] the government to eliminate their open source competition," suggesting that their popularity stems from fewer usage restrictions. He has also pointed to the inherent state censorship within Chinese AI models, a point often overlooked by those focused solely on their technical prowess. However, Sacks’s influence is diminished, having lost his formal advisory role. His stance favoring open AI has been increasingly overshadowed within the Trump orbit by a growing emphasis on government intervention, driven by national security concerns.
The prevailing view within this faction is that the sheer power of advanced AI models necessitates government control to mitigate potential national security threats. This perspective underpins the White House’s current review process, designed to vet AI models for security risks before their release. Dean Ball, a former Trump AI advisor now with OpenAI, has criticized this as a "de facto licensing regime for frontier AI." Ball speculated that Trump might resort to "soft power," potentially by making U.S. companies hesitant to adopt Chinese models like Kimi. This idea, however, was met with fierce opposition from Emil Michael, who, alongside Secretary of Defense Pete Hegseth, has been a key liaison with AI firms. Michael vehemently rejected the notion of clandestine government pressure, advocating instead for a transparent "democratic process, not some Deep State scheme."
Crucially, the ongoing debate largely sidesteps the question of how Chinese AI models achieved their current level of sophistication. For years, under both the Biden and the initial Trump administrations, efforts were made to restrict China’s access to cutting-edge semiconductor technology. While export controls have been loosened – a controversial move by Trump that allowed Nvidia to increase sales to China in exchange for a government revenue share – and reports of chip smuggling persist, China still faces computing power limitations. It remains unclear what specific hardware was used to train Kimi.
One plausible explanation for Kimi’s advancement is the practice of "distillation," where AI models are trained on the outputs of other AI models. OpenAI and Anthropic have consistently raised concerns about Chinese companies engaging in this practice and have sought government intervention. In April, the Trump administration responded with a series of measures aimed at curbing such exploitation of U.S. AI models.
Despite these efforts, Kimi remains freely available and is now nearly as capable as an Anthropic model that the U.S. government once deemed a national security threat, leading to its temporary shutdown. The recent public disagreements among Trump’s advisors underscore a shared awareness of the challenge posed by China’s AI advancements. However, the lack of consensus on the path forward reveals a deep ideological chasm, highlighting the complexity and internal divisions within the Trump campaign’s approach to the rapidly evolving landscape of artificial intelligence. The war of words among his advisors reflects a broader struggle to define America’s strategy in an AI-driven world, a struggle made more urgent by the potent challenge emanating from Beijing.
