Pentagon Confirms Grok AI Helped Strike 2,000 Iranian Targets

Pentagon Chief Digital and AI Officer Cameron Stanley disclosed in a sworn court filing that xAI's Grok chatbot assisted in launching 2,000 munitions that hit 2,000 distinct targets in just 96 hours. The statement was submitted in response to a lawsuit filed by the NAACP against xAI, which alleges that the company's Colossus 2 data center in Mississippi operated dozens of gas turbines without proper permits. In his testimony, Stanley emphasized that continued operation of the data center is critical to national security.

According to court documents, the operation was codenamed "Operation Epic Fury," the Trump administration's campaign against Iran. Stanley confirmed that the military used a specially built "Grok Gov" version of the chatbot, which was integrated with U.S. military targeting systems. This marks the first time a senior U.S. official has formally acknowledged in court documents the use of a commercial generative AI model in strikes against Iran, signaling a key step in the real-world deployment of commercial AI for kinetic targeting.

The disclosure has triggered urgent questions about auditability, legal liability, and vendor oversight in the military application of AI. Analysts note that while Grok's ability to process an average of more than 20 targets per hour is impressive, the direct involvement of a commercial AI model in lethal decision-making raises critical issues: ensuring algorithmic transparency, preventing bias, and establishing clear accountability. As AI becomes more deeply integrated into defense operations, establishing strict operational standards and legal frameworks is now urgent.