Read this post on Neil’s Substack: Getting Out of Control.
When National Economic Council Director Kevin Hassett called for an “FDA for AI” that would formally license frontier AI models, we warned against such an approach. The Trump administration subsequently walked back those comments. Yet, the U.S. may be on the road to something far worse.
ICYMI, on Friday, June 12, the federal government used export control authority to restrict access to Anthropic’s latest advanced models. Although a formal administration announcement is not yet public, Anthropic reports,
“The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.”
Follow-up reporting states that Commerce Secretary Howard Lutnick is imposing a licensing requirement on Anthropic’s models:
In a June 12 letter viewed by The New York Times, Mr. Lutnick told Mr. Amodei that a special license would be required for the company to distribute its Mythos and Fable 5 models “to all destinations worldwide,” as well as to share them with non-U.S. citizens.
This is not good! A leading U.S. AI company was forced to take down a product that millions were using based on non-public, unexplained concerns of a few government officials. This isn’t the red-tape risk of the FDA. It’s more like the FDA demanding, out of the blue and without explanation, that everyone stop drinking milk — if milk were 50% of last year’s stock market gains.
The administration has previously prioritized acceleration and winning the global AI race. In last summer’s “AI Action Plan,” the Trump administration sought to promote “the policy actions needed to sustain and enhance America’s AI dominance, and to ensure that unnecessarily burdensome requirements do not hamper private sector AI innovation.” It treated AI as a powerful general purpose technology with enormous and valuable private sector uses.
This latest USG action signals a change. It significantly escalates the centralization of control over advanced computation in our country. Like a recent Executive Order, the new export control demonstrates the rise of national security interests in regulating AI. Increasingly, the administration is treating AI like a weapon.
Various experts have questioned how the new “arbitrary, post-hoc system” of model-specific export controls blocking foreign nationals would even work. Anthropic chose the seeming only route to compliance by shutting down everyone’s access – and that may have been the government’s goal. Others have highlighted the important First Amendment concerns raised by these export controls, characterizing them as “a prior restraint on expression justified by speculative risk.”
It’s easy to be cynical about this story. No firm has done more to whip up a panic about frontier model capabilities than Anthropic. As former White House AI czar David Sacks has interpreted it, “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering. It is principally responsible for the state regulatory frenzy that is damaging the startup ecosystem.” In fact, just days before all this went down, Anthropic posted a new call for governments “to block or deter the deployment of models that pose a significant risk of catastrophic harm.”
No surprise then, that schadenfreude is on display across social media. Many have gleefully laughed about how Anthropic has all this coming, with common refrains of “FAFO” or “play stupid games, win stupid prizes.”
But even if you disagree with Anthropic’s regulatory strategy, this escalation of government intervention is nothing to celebrate. It is horrible for the broader AI ecosystem. Continued arbitrary, unexplained deployment of export control authority will make companies slow-walk new models, depriving the public of powerful new tools. Every AI model, like all software before it, will have vulnerabilities that require patching. The US government should not hang a Sword of Damocles over every lab’s head, with no indication when it might drop or why.
Indeed, it’s plausible that this development actually advances Anthropic’s strategy. They may see this as a partial win. There is now established precedent for the USG forcibly blocking a model release and requiring a license. When the (unsustainable) export controls fall away, Anthropic will no doubt meekly suggest as a replacement the FDA-style pre-release government licensing regime they’ve called for (and we’ve warned about) all along.
This episode yet again shows why Congress must act. We need a balanced statutory framework for frontier model safety, rooted in the rule of law, with clear standards and transparent procedures. Civilian authorities must direct this process; it must not be co-opted by the military-industrial complex. America’s AI leadership will diminish if our government continues the ad-hoc and myopic approach to AI policy recently on display.