The Pentagon says it has stopped using Anthropic’s AI tools, months after moving to blacklist the company. The BBC also reports that people familiar with defence operations said Claude was still in use as recently as last week, including through the military’s Maven data platform.
Anthropic Watch analysis
What happened
A US defence department official told the BBC that the department “has ceased the use of Anthropic products”. The BBC says the Pentagon had set a late-August deadline to phase out Claude after designating Anthropic a national security supply-chain risk.
The BBC’s sources said Claude remained in use in recent weeks for research, analysis and intelligence work. Several described it as integrated into Maven, a Palantir-operated platform used to organise intelligence data. Anthropic declined to comment on the Pentagon’s statement, the BBC reports. Read the BBC’s report.
Why it matters
This is not just a dispute over which model gets a government contract. The BBC’s account describes an AI tool embedded in a wider intelligence system, where replacing it may be more involved than switching off a standalone service. The reported uses, including analysing imagery and preparing military briefings, make the gap between an announced phase-out and actual removal consequential.
The distinction is important: the department says use has stopped; the BBC’s sources describe use continuing into the previous week. The report does not establish precisely when, or how completely, that transition happened.
Our read
The key question is whether the Pentagon’s announcement marks a completed removal or the end of a drawn-out process. The BBC’s account makes a strong case that integration can outlast a policy decision. It also reports that Anthropic challenged the supply-chain designation in court, while the department has signed contracts with other AI companies. A technology ban is one thing on paper; unwinding it across a working system is quite another.
What to watch
- Whether the Pentagon clarifies when Claude was fully removed and from which systems.
- Whether Anthropic or the department provides a substantive response to the reported continued use.
- How Maven and other defence systems replace Claude, and whether other models take its place.
Discussion spark: If a government orders an AI tool removed but it remains embedded in operational systems, should the deadline be judged by the policy announcement or by when the tool is actually gone?
Sources and evidence
- Pentagon no longer using Anthropic AI tools, BBC told – bbc.com (5 October 2026, 16:13 UTC)
Anthropic Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by Anthropic.