Skip to content
Submit incident
AII-1001-13781
AllegedThe record is a lawsuit, complaint or accusation that has not been decided.

US Military Used Anthropic AI in Iran Strike Despite Trump Ban

Reportedly, the US military used Anthropic's Claude AI in an Iran strike despite Trump ordering federal agencies to stop using Anthropic's technology, which the company had refused to release for unrestricted military use. The Pentagon classified Anthropic as a supply chain risk, and Anthropic announced plans to challenge the designation in court.

28 February 2026 · Anthropic · Security and cyberattacks
28 February 2026
AII-1001-137814 sources ↓
Quick facts
Category
Security and cyberattacks
Country
United States
Source
Фактор
Date
28 February 2026
Drafted by
Raidu Incident Detection

What happened

The US military reportedly used Anthropic's Claude AI system in a strike against Iran, despite President Donald Trump ordering the US government to cease using Anthropic's technology.

Trump had directed agencies to halt use of Anthropic's models, citing national security concerns because the company refused to permit unconditional military use and opposed mass surveillance and fully autonomous weapons systems.

The Pentagon announced it would classify Anthropic as a supply chain risk, and Anthropic indicated it would challenge this designation in court.

Reported impact

The Pentagon's supply chain risk designation could prevent tens of thousands of military contractors from using Anthropic's AI systems, threatening the company's business with the government.

Anthropic CEO Dario Amodei stated that no pressure or punishment from the Department of Defense would change the company's position on mass surveillance or fully autonomous weapons.

Index analysis

The dispute reflects tensions between AI companies and the US military over safeguards for autonomous weapons and surveillance capabilities, with Anthropic maintaining constraints that Trump's administration sought to remove.

Governance controls

Anthropic refused to permit unconditional military use of its models and opposed mass surveillance and fully autonomous weapons systems.

OpenAI announced agreement with the Pentagon to use its models in classified networks with 'technical protective mechanisms' and requirements that humans make decisions when force is used.

Classification

Category
Security and cyberattacks
Industry
Government, security and defence
Country
United States
Harm status
Actual harm
Intent
Unknown
Timing
Unknown
Harm types
Public interest, Other
Harmed
Government
Consequences
Litigation
AI systems
Claude

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Фактор
Пентагонът сключи договор с OpenAI - Фактор
28 February 2026
frognews.bg
Anthropic vs. Тръмп: В САЩ се развихря AI война
28 February 2026
Bloomberg
След сблъсъка с Anthropic: OpenAI допуска Пентагона до своя AI модел
28 February 2026
168 Часа | By 24 часа онлайн
Военните на САЩ използвали забранен от Тръмп изкуствен интелект при атаката срещу Иран
2 March 2026