Back to News
PolicyAI Understanding briefing

US appeals court upholds Pentagon ban on Anthropic's Claude model

A federal appeals court affirmed the Pentagon’s decision to bar Anthropic from its defense supply chain after the AI firm refused to lift safeguards on its Claude model.

4 min readRead the linked source
Source-page capture accompanying US appeals court upholds Pentagon ban on Anthropic's Claude model
Source referenceSource recorded
Publisher
24newshd.tv
Source link
24newshd.tvhttps://www.24newshd.tv/26-Sep-2026/us-court-sides-pentagon-anthropic-ai-ban?version=amp
Source type
Linked source — primary-source status has not been established.
Also cited

Story last revised

ContextUnderstand this in 60 seconds

Start here

Key terms

AI Governance
Policies, standards, and oversight mechanisms that guide how AI is developed and used in society.
Weight
A learned numeric value that scales signals passing through a neural network.
Test yourselfAI Ethics Quiz

What changed since publication

  1. First published
  2. The AFP report adds new details to the ongoing story, including the court’s reference to an Anthropic executive’s objection to Palantir’s use of Claude during the operation that captured Venezuelan President Nicolás Maduro, direct quotes from Under Secretary of Defense Emil Michael, and the contrast with a separate California federal court decision that found a broader Trump‑era ban illegal. These specifics deepen understanding of the legal rationale and the broader policy conflict surrounding AI safety safeguards.
  3. The appeals court affirmed the Pentagon’s supply‑chain risk designation against Anthropic, keeping the company barred from military contracts after it refused to lift safeguards on its Claude model.
  4. The appeals court affirmed the Pentagon’s ban on Anthropic, citing the company’s refusal to lift safety restrictions on Claude and an internal objection to its use by Palantir during the Maduro capture operation, thereby solidifying the government’s authority to blacklist AI vendors for national‑security reasons.

What happened

The 2nd U.S. Circuit Court of Appeals affirmed the Pentagon’s designation of Anthropic as a supply‑chain risk, keeping the company out of all Department of Defense contracts. The ruling follows Anthropic’s refusal in February to remove restrictions that would allow its Claude model to be used for fully autonomous lethal weapons or mass domestic surveillance. The court’s opinion noted that an Anthropic executive had raised concerns about Claude’s use by a Pentagon contractor, Palantir, during the operation that captured Venezuelan President Nicolás Maduro on Jan. 3, prompting “alarm” among senior defense officials.

On Friday, the 2nd U.S. Circuit Court of Appeals, in a decision authored by Judge Gregory Katsas, upheld the Pentagon’s March supply‑chain risk designation against Anthropic. The designation cancelled existing military contracts and prohibited other defense contractors from using Anthropic’s Claude model.

The court noted that Anthropic’s refusal to remove safeguards on Claude—specifically those preventing fully autonomous lethal weapon use and mass surveillance—constituted a material risk to warfighter safety. An internal objection by an Anthropic executive regarding Claude’s deployment by Palantir during the Maduro capture operation was cited as a catalyst for the Pentagon’s concerns.

Under Secretary of Defense Emil Michael’s statements, quoted in the opinion, emphasized that the presence of Anthropic’s safeguards could cause software failures or other disastrous outcomes, endangering service members.

Anthropic maintains that it cannot retroactively modify models already delivered, but the court found that the company still controls the behavior of each new version, obligating the military to stay current with updates.

Source details: 24newshd.tv ↗

Why it matters

The decision reinforces the U.S. government’s authority to blacklist AI vendors that do not align with military risk‑acceptance standards, setting a precedent for future AI procurement and export‑control policies. By keeping Anthropic’s technology out of classified systems, the ruling may shape how other AI firms negotiate safety constraints with defense customers. It also adds legal to the Pentagon’s broader effort to secure its AI supply chain amid concerns about model reliability, ethical use, and potential sabotage. The case highlights tension between commercial AI development—Anthropic is valued near $1 trillion and preparing for a major IPO—and national‑security imperatives, potentially influencing investor sentiment and prompting other firms to pre‑emptively adopt stricter safeguards.

The ruling clarifies that the President and Secretary of Defense have ultimate authority to balance AI risks against operational needs, reinforcing executive control over AI procurement.

By affirming the blacklist, the decision may deter other AI firms from offering unmodified models to the defense sector, encouraging pre‑emptive safety features and transparent risk assessments.

The case underscores the growing intersection of and national security, signaling that commercial AI strategies must account for potential regulatory constraints before seeking government contracts.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Interactive Concept Check+10 Points
AI Ethics Quiz

Impossibility results in algorithmic fairness (e.g. Kleinberg et al., Chouldechova) show what?

What to watch next

Future litigation could arise from the separate California federal court ruling that declared a broader Trump‑era ban on Anthropic illegal, creating a possible legal clash. Anthropic’s upcoming IPO and its negotiations with the White House will be closely monitored for any concessions on model controls. Additionally, the Pentagon may issue further guidance on AI risk assessments, affecting contracts with other vendors such as OpenAI, Google, and Microsoft.

Potential appeals or conflicting rulings from other federal courts could create legal uncertainty for the Pentagon’s supply‑chain policy.

Anthropic’s upcoming IPO may be affected by investor perception of regulatory risk and the company’s stance on model safeguards.

The Pentagon is expected to release further guidance on AI risk labeling, which could expand the blacklist to additional vendors or tighten existing criteria.

Related guides & quizzes

AI EthicsAI Models ExplainedTest what you know — try a free AI quizLook up an AI term in our glossaryFollow the AI regulation tracker

Updates and corrections

This canonical story is updated in place when the developing event materially changes. Its URL and original publication date never change.

  • The appeals court affirmed the Pentagon’s ban on Anthropic, citing the company’s refusal to lift safety restrictions on Claude and an internal objection to its use by Palantir during the Maduro capture operation, thereby solidifying the government’s authority to blacklist AI vendors for national‑security reasons.
  • The appeals court affirmed the Pentagon’s supply‑chain risk designation against Anthropic, keeping the company barred from military contracts after it refused to lift safeguards on its Claude model.
  • The AFP report adds new details to the ongoing story, including the court’s reference to an Anthropic executive’s objection to Palantir’s use of Claude during the operation that captured Venezuelan President Nicolás Maduro, direct quotes from Under Secretary of Defense Emil Michael, and the contrast with a separate California federal court decision that found a broader Trump‑era ban illegal. These specifics deepen understanding of the legal rationale and the broader policy conflict surrounding AI safety safeguards.
See the public corrections log
Found this useful?