A sudden decision by the Donald Trump administration to limit access for all foreign nationals to Anthropic’s latest artificial intelligence models has sparked confusion and concern inside the company.
The restrictions, which affect the Fable 5 and Mythos 5 systems, were imposed over what officials described as national security concerns.
The move reportedly came with little warning. Anthropic executives were informed on Friday that access to the models would need to be suspended for foreign users, The New York Times reported.
Nearly a week after the surprise order, many workers say they still have little clarity on why the restrictions were imposed. The sudden warning sparked panic across the company, with internal discussions quickly filling with questions about the nature of the threat and whether the decision could affect Anthropic’s plans to go public later this year.
The controversy began last week when Anthropic executives were informed by the White House that they had less than 90 minutes to remove their newly launched AI models. The message triggered a rush inside the company to prepare customers for possible disruptions.
According to the NYT report, explanations from management changed several times. Some workers were initially told that foreign companies could gain access to the models, while others later heard that a significant security vulnerability had been discovered.
In a message reviewed by The New York Times, one worker asked, “What are you telling your clients?” while another questioned, “Does anyone know what to believe?” A third employee admitted, “I don’t understand what the issue is.”
Despite ongoing meetings between Anthropic Chief Executive Dario Amodei and Trump administration officials, no resolution has been reached. Anthropic said on Monday that it plans to continue talks with government representatives and remains committed to working alongside the administration.
The dispute marks the second major clash between Anthropic and the Trump administration this year over its powerful technologies. Earlier, the company became embroiled in a disagreement with the Pentagon over the use of AI in warfare, leading Defense Secretary Pete Hegseth to label Anthropic a “supply chain risk”.
Tensions increased further after Anthropic unveiled Mythos, an advanced AI model capable of identifying software vulnerabilities. Later, Anthropic released Fable 5, a straitjacketed version of Mythos designed with additional guardrails and intended for wider use.
The situation escalated after researchers at Amazon identified what they viewed as a security shortcoming in Fable 5. Their preliminary findings suggested the model could be prompted to reveal flaws in vulnerable software code. Government officials reportedly reviewed the research and considered the implications serious.
Inside Anthropic, however, uncertainty remains high. Employees continue to debate the reasons behind the government’s actions and what they could mean for the company’s future. “At what point does this just feel like they don’t want us to exist?” one employee asked on Tuesday.
Support for Anthropic has emerged from the cybersecurity experts, who stated that the Trump administration was unfairly targeting the company. More than 150 experts signed an open letter urging officials to lift the restrictions.
