The Trump administration’s decision to force Anthropic to pull two of its newest AI models offline last Friday has ignited a complex debate about AI policy, digital sovereignty, and the murky line between national security and political retribution. The order, which cited unspecified national security concerns, has left the AI industry and cybersecurity experts questioning the administration’s motives and the broader implications for the sector.
The specifics of the order
According to reports, the U.S. government sent a letter to Anthropic on Friday demanding that it ensure its two newest models — Fable 5 and Mythos 5 — could not be used by any foreign nationals. The administration provided no public details on the specific national security risks, leading Anthropic to conclude it had to pull the models entirely. As TechCrunch’s Rebecca Bellan noted on the Equity podcast, “We don’t know when someone’s a foreign national. A lot of our own employees are foreigners.”
Also read: Using AI for financial advice? Proceed with caution
The political backdrop
The order came at a time when the Trump administration was also reportedly trying to negotiate a treaty related to the war in Iran, leading some to view the timing as a distraction. More significantly, the move appears to be part of a strained relationship between the administration and Anthropic. As TechCrunch’s Sean O’Kane observed, “Anthropic has not had the best relationship with the Trump administration in a way that stands apart from the other leading AI labs.” This tension is further highlighted by an ongoing lawsuit between the government and Anthropic, where the company was labeled a supply chain risk.
Who benefits from the crackdown?
The question of who benefits from this regulatory action is central to the debate. On one hand, Anthropic’s rivals — such as OpenAI and Google DeepMind — could see a temporary advantage. If the administration is less inclined to take similar action against them, they may have more freedom to release powerful models. However, cybersecurity experts have pushed back, with many signing an open letter asking the president to revoke the order. They argue that pulling advanced cybersecurity capabilities from U.S. network defenders is itself a dangerous move. As Rebecca Bellan summarized, “They say it’s actually dangerous to have to pull these advanced cybersecurity capabilities from network defenders in the U.S.”
Also read: Snap Spins Off AI Video Team Into New Company, Dotmo, to Cut Costs
Anthropic’s own messaging paradox
Anthropic’s public stance on AI safety may have contributed to the scrutiny it now faces. The company has long warned about the dangers of powerful AI, even as it releases increasingly capable models. This apparent contradiction was noted by TechCrunch’s Anthony Ha, who pointed out that Anthropic’s marketing of Mythos as “the most incredibly powerful model ever, it’s too dangerous to release to the public” invites an “incredible level of scrutiny.” This duality may have made the company a more vulnerable target for regulatory action.
Ultimately, the situation may prove to be a double-edged sword for Anthropic. While the forced removal of its models is a significant setback, the resulting publicity could boost the company’s profile. As Rebecca Bellan noted, “Everyone loves a bad boy.” Downloads of Anthropic’s Claude assistant surged after a previous dispute with the administration, suggesting that controversy can sometimes be a powerful marketing tool.
Frequently Asked Questions
What specific AI models did Anthropic have to pull?
Anthropic was forced to pull two models: Fable 5, which was publicly available, and Mythos 5, which was available to existing Mythos users.
Why did Amazon researchers reportedly trigger this order?
Reports indicate that Amazon researchers found a way to bypass Fable 5’s guardrails, and Amazon CEO Andy Jassy raised these concerns with the White House, leading to the export control order.
Has the Trump administration taken similar action against other AI companies?
Not yet. Experts note that Anthropic has a uniquely strained relationship with the administration, making it unclear if other AI labs face similar risks.
