AI & Singularity
Guardian Deep-Dive on Anthropic-Pentagon Standoff
The Guardian has published the most comprehensive analysis yet of Anthropic’s standoff with the Pentagon, exploring how the AI safety company became the central actor in a debate over military AI use that could shape the future of AGI development.
The Guardian explores how Anthropic became the central actor in AI ethics battle.
The Origins
The feature details Anthropic’s founding by Dario and Daniela Amodei after leaving OpenAI in 2021, highlighting the company’s deep ties to the effective altruism movement through investors and leadership connections.
The Contradictions
The Guardian piece explores inherent tensions in Anthropic’s position:
- Safety-first mission vs. classified military work
- Partnership with Palantir for defense applications
- Claude was the first model permitted for classified military use
- Military using Claude via Palantir’s Maven system to determine bombing targets in Iran
The “Double Black Box” Problem
A key insight: the military doesn’t know how Claude works, and Anthropic doesn’t have perfect visibility into classified use. This creates accountability gaps that the Pentagon standoff has brought into sharp focus.
As one Hugging Face ethics scientist quoted in the piece observed: “It’s not that they don’t want to kill people. It’s that they want to make sure to kill the right people. And who the right people are is decided by the government.”
Current Status
Anthropic reportedly reopened negotiations with the Department of War in recent days. CEO Dario Amodei apologized for the tone of a leaked internal memo, clarifying that the company’s objections specifically concern autonomous weapons and mass surveillance ” — not operational decision-making.
Source: The Guardian, March 9, 2026
