ai
Anthropic's Claude chatbot questioned its role in U.S. military targeting system
The Atlantic reported that Claude, an Anthropic large language model, generated responses critical of its use in the Maven Smart System. The system fuses intelligence data to support battlefield targeting decisions.
app.buzzsumo.comAudio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
Anthropic's Claude large language model produced written responses that described its participation in a U.S. military targeting platform as troubling. The author posed questions to Claude in Amsterdam two months before the June 2026 publication date. Claude stated that generating targeting coordinates for airstrikes was far from its intended purpose of being helpful, harmless, and honest.
Military platform details Claude is integrated into the Maven Smart System, which combines satellite imagery, drone feeds, and communications intercepts. Officers can use the system to locate enemy units, select weapons, and plan attack angles in minutes rather than hours.
A February strike using the system hit an elementary school in Minab, Iran, killing about 170 people. Investigators were examining whether outdated satellite imagery contributed to the error.
Company dispute Anthropic is engaged in litigation against the Pentagon over its March designation as a supply-chain risk. The designation could restrict the company's government contracts. Claude stated that Anthropic drew a line against certain military uses and that the line was appropriate.
The company has sought to prevent its products from being used in autonomous lethal-weapons systems or for mass surveillance of Americans.
