Dan Wang & Julian Gewirtz
A globally recognisable tech executive, high-spirited from preparing for a public offering, offers imprudent remarks criticising the government. The state strikes back harder than anyone expects. Overnight, the bargain between a skyrocketing sector of the economy and the government is shattered.
If you think this story could be about Anthropic, you’re only half right. In 2020, Alibaba co-founder Jack Ma found himself in the doghouse after he publicly rebuked Chinese regulators. Citing regulatory concerns, authorities cancelled the public offering of Ant Group and subsequently unleashed a regulatory storm that left few Chinese tech companies untouched.
The US government is skating close to its own Jack Ma moment, when a government wounds a tech leader seemingly out of spite. Self-destructive American actions, not Chinese competition, may be the most significant threat to the evolution of artificial intelligence for years to come.
On June 9, Anthropic released Fable 5, an adapted version of its powerful Mythos model, which has an incredible capacity to find vulnerabilities in software. Anthropic’s chief executive, Dario Amodei, even said that companies using Mythos called it a “superweapon.” Three days later, the US government issued an export control directive blocking use of Fable 5 by foreigners and noncitizens, which prompted Anthropic to disable all access. A tangle of explanations behind this directive has been reported, including the risk of “jailbreak” and access by foreign adversaries. The government later permitted Anthropic to restore some users’ access to a version of Mythos, though negotiations are still underway about Fable 5.
Over the past decade, the US government has used export controls to deal crippling blows to Chinese technology champions. The action against Anthropic upended this logic by turning this policy instrument against American companies, purportedly to exert the government's grip over increasingly slippery AI models. The challenges of regulating AI at the frontier have now reportedly moved the government to ask Anthropic’s chief competitor, OpenAI, to limit the users for its own next model.
We often think of AI as a race between the US and China. Instead, we are seeing the emergence of an even more acute competition between the public power of governments and the private power of ambitious companies. Both countries are struggling to determine whether their frontier AI companies are national champions or national security threats. The US government needs to strike a better balance between ambition and control, lest it irrevocably damage its relationship with these companies and America’s long-term technological edge.
The second Trump administration has careened from extreme to extreme on AI policy. It came into office preaching the virtues of a hands-off approach. Its stance shifted in March: The Pentagon designated Anthropic a supply chain risk after the company protested the use of its AI models in autonomous weapons and domestic mass surveillance. Reports about Mythos’s capabilities the following month appear to have shocked the administration.
The Chinese government, meanwhile, has been overseeing its AI labs for considerably longer. It requires security assessments and tests models for their ability to anticipate Beijing’s political sensitivities. The state has blocked Chinese companies from purchasing advanced Nvidia chips to favour the domestic chip industry, restricted overseas travel by certain AI researchers, and ordered the unwinding of Meta's acquisition of the Chinese AI agent startup, Manus.
Unlike in the US, in China, no one questions who will ultimately win the struggle between state and corporate power. But the effects of Beijing's regulatory storm remain evident. Alibaba’s stock has lost around two-thirds of its value, venture capital funding has collapsed, and conversations with Chinese AI labs include bitter complaints about government overreach.
The US is not likely to suffer to the same degree, and the companies and government officials will likely reach an understanding. But Washington and American AI labs need to treat each other with greater sincerity and seriousness going forward.
First, the heads of AI labs need to stop their doom trolling. It makes little sense to make panicked claims about the destructive potential of AI without a plan to work with the government to address those risks. Second, the US government needs to realise that the stakes of AI are far too high to allow a breakdown of trust. Third, America’s government and its AI labs need to update their understanding of Chinese AI.
The AI chaos of recent weeks is self-defeating. For the US to win the AI future, it needs to do better at avoiding the missteps of Beijing.