Claude On Battlefield? Iran Used AI To Track US Warships, Yemen To Develop Missiles, Anthropic Reveals

News world Claude On Battlefield? Iran Used AI To Track US Warships, Yemen To Develop Missiles, Anthropic Reveals

Claude On Battlefield? Iran Used AI To Track US Warships, Yemen To Develop Missiles, Anthropic Reveals

Published By :

,

Last Updated: September 11, 2026, 12:22 IST

An Iran-linked group used Anthropic's Claude AI to target US naval forces, collecting intelligence and compiling "targeting handbooks".

3 Quick Answers

An April 8, ceasefire ended nearly 40 days of intense bombardment, but sporadic attacks mostly near the strategic Strait of Hormuz have dampened the prospect of a full-fledged peace agreement. (File Image source: AFP)

An Iran-linked group used Anthropic’s Claude artificial intelligence model to gather intelligence on US naval forces, track their movements and compile material that could be used for military targeting, according to a new report by the AI company.

Anthropic said it had identified and disrupted six operations spanning Iran, Yemen, China and Russia in which its AI model was used for activities ranging from intelligence collection and weapons development to procurement.

The cases highlight how generative AI is being incorporated into military and defence-related work that traditionally required specialised teams of engineers and analysts.

Iran-Linked Operators Turned Claude Into US Navy Targeting Tool

In the Iran-linked case, operators used Claude to collect and analyse publicly available information about US naval assets in the region.

The material was then organised into what Anthropic described as “targeting handbooks", including information that could help identify and track naval positions.

Also Read: ‘Don’t Believe In The Word Regret’: Trump Says He’d Change Nothing About US’ Iran War Strategy

“We identified and disrupted an Iran-nexus threat actor that used Claude to collect and analyse publicly accessible data to develop targeting recommendations against US naval forces in the region," Anthropic said.

The information assembled by the operators included names of US personnel extracted from captions accompanying publicly available military photographs, ship and aircraft transponder identifiers, scripts for querying commercial satellite imagery and lists of websites that publicly displayed information about US naval movements.

‘Claude Used To Hunt Weakness In Systemes Used Aboard US Ships’

The group also turned to Claude for technical research into potential weaknesses in systems used aboard ships. Anthropic said this included vulnerability research involving maritime satellite terminals, Cisco communications equipment and industrial control products.

“The threat actor used Claude to compile targeting handbooks, through a Python pipeline the threat actor built with Claude’s assistance, to identify and track naval positions based on open-source information. The compiled material included a roster of US personnel scraped from captions on public military photographs; publicly accessible ship and aircraft transponder identifiers; commercial satellite-imagery query scripts; and an inventory of public websites that exposed US naval movements. The threat actor also directed Claude to compile vulnerability research on shipboard systems, cataloguing known CVEs in maritime VSAT terminals, Cisco communications equipment, and industrial control products," the report stated.

The disclosure comes amid heightened military confrontation involving Iran and US forces in the region.

Also Read: Joke Or Subtle Hint? Trump Says He’s Getting A ‘Little Bit Tired’ Of Politics As Iran War Escalates

Iranian forces have launched attacks targeting US naval assets, including the guided-missile destroyers USS Delbert D Black and USS John Paul Jones. Iran also targeted Naval Support Activity Bahrain, which houses the headquarters of the US Navy’s Fifth Fleet.

US officials said the warships were able to evade the attacks without casualties, while the strikes on the Bahrain facility caused damage.

Yemen’s Houthis Use Claude For Missile Development

The report also details a separate operation involving a Yemen-based weapons engineering cell that used Claude to work on missile and guided-rocket programmes.

Anthropic said the group was developing several missile systems, including a multi-stage ballistic missile that it claimed would have a range exceeding 2,000 km, as well as an “R2000" system that included a variant featuring a hypersonic glide vehicle.

The operators used Claude Code to develop software involved in guiding and controlling flying vehicles. Anthropic said the model was effectively being used “in place of human software engineers".

The work involved combining an open-source autopilot with a flight computer and writing software for navigation, control and position estimation. The group also used the AI model for tasks such as adjusting control parameters, firmware development and flight simulations.

The operators appeared to divide the work among several Claude instances, with different sessions assigned to coding, research and reviewing software generated elsewhere.

Also Read: Saudi Crown Prince MBS Asked Trump Twice For Strikes On Houthis, US President Declined: Report

Anthropic said it found no evidence that the group had successfully deployed an operational weapon. However, the group did conduct a test of a guided rocket, which appeared to fail. The operators returned to Claude shortly afterwards to try to determine the cause of the failure.

The company said its safety systems stopped a significant number of requests but acknowledged that they did not prevent all attempts.

The operators also attempted to hide the nature of their activities, according to Anthropic, by distributing their work across separate sessions and disguising the intended purpose of individual requests.

The company said it subsequently banned accounts connected to the operation and shared relevant information with public- and private-sector partners. The group had also built an offline simulation system that could operate without Claude or conventional engineering programmes such as MATLAB.

Claude’s Role In Organising Khamenei’s Funeral

Anthropic also identified another Iran-linked use of Claude involving the Islamic Culture and Communications Organization.

According to the company, an account associated with the organisation uploaded detailed organisational blueprints concerning Ayatollah Ali Khamenei’s funeral and succession arrangements.

The cases come as AI companies face growing pressure to prevent their models from being exploited for military, intelligence and weapons-related purposes, while also grappling with the difficulty of distinguishing legitimate technical research from activities that could facilitate real-world harm.

News18 Newsletter Handpicked stories, in your inbox A newsletter with the best of our journalism submit

Quick Answers Powered by Ask News18 How did an Iran-linked group use Anthropic's Claude AI to target US naval forces? An Iran-linked threat actor used Claude to collect and analyze open-source data to compile "targeting handbooks" and track US naval positions. The compiled information included names of US personnel scraped from public military photo captions, ship and aircraft transponder identifiers, commercial satellite-imagery query scripts, and public websites exposing naval movements. The group also used Claude to research vulnerabilities in shipboard systems, including maritime satellite terminals, Cisco communications equipment, and industrial control products. How did Yemen-based operators use Claude AI for missile development? A Yemen-based weapons engineering cell used Claude to work on missile and guided-rocket programs, including a multi-stage ballistic missile and an "R2000" system featuring a hypersonic glide vehicle. The operators used Claude Code to write software for navigation, control, and position estimation, effectively using the AI in place of human software engineers. They also used the model for firmware development, adjusting control parameters, and running flight simulations, later using Claude to analyze the failure of a guided rocket test. What did Anthropic discover regarding Claude's involvement with Ayatollah Ali Khamenei's succession? Anthropic identified an Iran-linked account associated with the Islamic Culture and Communications Organization that used Claude. This account uploaded detailed organizational blueprints concerning Ayatollah Ali Khamenei's funeral and succession arrangements to the AI model. Powered by Ask News18

About the Author Prisha Vibhavari Prisha is the Chief Sub-Editor at News18.com, with more than 10 years of experience in national and international news. She specialises in editorial leadership, sharp news judgment, and high-impact st... Read More

ADD YOUR COMMENT

Location : Washington D.C., United States of America (USA)

First Published: September 11, 2026, 12:12 IST

Disclaimer: Comments reflect users’ views, not News18’s. Please keep discussions respectful and constructive. Abusive, defamatory, or illegal comments will be removed. News18 may disable any comment at its discretion. By posting, you agree to our Terms of Use and Privacy Policy Loading comments...