Anthropic Employees Accuse Trump Administration of Targeting Them
The New York Times
Trump officials banned Anthropic's AI models from government use, and Anthropic staff say they're being singled out. Over 150 cybersecurity experts just signed a letter demanding the ban get reversed.
Based on reporting by The New York Times — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Something's off when the company most cybersecurity researchers point to as the safety-conscious one in AI ends up on the wrong side of a federal ban. That's the situation Anthropic finds itself in now, after the Trump administration restricted use of its Claude models across government agencies. Anthropic employees are pushing back hard, arguing the move looks less like a policy decision and more like a targeted swipe at a company that has spent years publicly worrying about AI risk, sometimes to the annoyance of officials who'd rather move fast.
The backlash didn't stay internal. More than 150 cybersecurity experts signed an open letter this week urging the administration to lift the restrictions, a number that's hard to dismiss as a fringe complaint. These are people who work with model security day in and day out, and their argument boils down to this: banning one of the more rigorously tested AI labs doesn't make federal systems safer, it just narrows the field of tools available to people trying to secure them.
What makes this particularly awkward for the administration is Anthropic's reputation. The company has built its entire public identity around caution, releasing research on model interpretability and pushing for guardrails well before competitors did. If there's a lab you'd expect a security-minded ban to spare, it's this one. Employees inside Anthropic have framed the restriction as retaliation, pointing to the company's outspoken stance on AI regulation as the likely trigger rather than any actual security flaw in Claude.
Anthropic hasn't detailed a legal response yet, and the administration hasn't offered much public justification beyond general security language. But the optics are rough: a government citing safety concerns to sideline the AI company most associated with taking safety seriously, while hundreds of independent experts say the ban doesn't hold up to scrutiny.
My take — AI-written commentary, not fact-checked reporting
I've read a lot of vague government justifications for tech bans, and this one smells like politics wearing a security costume. Anthropic has been the industry's most annoying conscience on AI risk, and punishing the company that keeps raising uncomfortable questions is exactly the wrong incentive to set for everyone else. If 150-plus security researchers think the ban doesn't check out, that's not a rounding error, that's a consensus worth listening to.
Read more about this at: The New York Times
