The claim
“The contract language proposed by the Department of War does not effectively prohibit the use of Anthropic's AI for mass surveillance or autonomous weapons.”
Anthropic made the claim on 4 October 2026 in an ABC News article about the Pentagon contract.
Our verdict
Unverified
We could not verify whether the proposed contract language effectively prohibits the use of Anthropic's AI for mass surveillance or autonomous weapons because the actual contract text is not publicly available.
Know something that settles it? Send us evidenceWhat we found
Anthropic stated publicly that the contract language made no progress on its demands to ban mass surveillance and fully autonomous weapons, as reported by ABC News. A CNBC report and a federal court order confirm that negotiations and litigation took place between the Department of Defense and Anthropic. A teleSUR English article adds that the Pentagon demanded the removal of safety restrictions on those uses, which Anthropic refused. However, none of the available sources provide the actual proposed contract wording, so we cannot independently assess whether the language effectively prohibits those activities.
The evidence
1 contradict and 1 support the claim, of 5 sources.
- Supports the claim
A CNBC report on 25 September 2026 confirms that the Department of Defense and Anthropic engaged in negotiations over the use of Claude AI models before the Pentagon designated Anthropic as a supply chain risk.
cnbc.com· U.S. appeals court upholds Pentagon designation of Anthropic as supply chain riskSeptember 25, 2026
- Background
An ABC News article reports that Anthropic publicly stated it conditioned its agreement with the Department of War on prohibitions against mass surveillance and fully autonomous weapons.
abcnews.com· Trump orders US government to cut ties with Anthropic; Hegseth declares supply chain 'risk' - Contradicts the claim
A teleSUR English article on 26 September 2026 says the Pentagon demanded Anthropic remove safety restrictions on autonomous weapons and mass surveillance, which Anthropic refused.
telesurenglish.net· Anthropic Pentagon AI Ban: 5 Critical Risks in the Court’s Security Ruling - teleSUR EnglishSeptember 26, 2026
- Background
A federal court order dated 27 August 2026 confirms litigation between Anthropic and the Department of War over the supply chain risk designation, noting the dispute involved Anthropic's public criticism of the Department's views on AI use.
104.18.4.58· united states district courtAugust 27, 2026
- Background
Anthropic's March 2025 submission to the Office of Science and Technology Policy outlines its recommendations for an AI action plan, providing background on its policy positions.
160.79.104.10· Anthropic Response to OSTP RFI (March 2025) - Final Submission (v3)March 6, 2025
What we could not confirm
- The exact wording of the Department of War's proposed contract language.
- Whether that language effectively prohibits the use of Anthropic's AI for mass surveillance or autonomous weapons.
How we checked
We reviewed public statements from Anthropic, news reports from CNBC, ABC News and teleSUR English, a federal court order, and Anthropic's own policy submission to the US government. We reviewed 5 relevant sources.
Help us check this claim
We could not find enough public evidence to confirm or refute this claim. If you know of an official record, dataset, transcript or report that settles it, send us the link. We review every submission and update this fact check when the evidence changes the picture.
Measured Take updates fact checks when new evidence emerges. Corrections are noted on the page.
All fact checks