St. Claude's halo: does Anthropic's human-centric image hold up on a closer look?

St. Claude's halo is shining bright. Refused the Pentagon. Teamed up with the Pope. Their image has never looked better, but does it hold up when you look more closely?

The name "Anthropic" comes from *anthropos*, Greek for human. The company is built around the idea of human-centric AI. If you read the headlines, the last 90 days look like the mission delivered.

The Pentagon refusal in February:In Anthropic's February 2026 statement on autonomous weapons, CEO Dario Amodei wrote: "Frontier AI systems are simply not reliable enough to power fully autonomous weapons." In the same statement, he called partially autonomous weapons "vital to the defense of democracy," and said fully autonomous ones "may prove critical." That reads as a "not yet," with the moral question left open.

Claude Gov:Anthropic builds custom models for U.S. national-security customers. Classified networks. Intelligence analysis. Intelligence and defense documents. Cybersecurity data analysis. Anthropic's own words, from their Claude Gov product page.

Palantir:Since November 2024, Claude has been operational inside Palantir's defense platform, in the IL6 environment reserved for highly sensitive national-security systems. Per the joint Palantir-AWS-Anthropic announcement.

What Palantir stands for. In April 2026, Palantir posted a 22-point public manifesto condensing CEO Alex Karp's book "The Technological Republic". It pushes for AI weapons. It declares some cultures "dysfunctional and regressive."

Together they do not add up to the picture the brand tells.

Don't get blinded by the perfectly staged marketing.

Ask where your own values lie. Then ask whether the AI you depend on shares them.