Pentagon used xAI Grok model in Iran strikes, filing says

A top Pentagon artificial intelligence official stated that the Grok chatbot developed by xAI supported the identification of more than 2,000 targets during U.S. operations in Iran. The disclosure came in a sworn court filing defending xAI against a lawsuit alleging the company violated the Clean Air Act by operating dozens of gas turbines without required permits.

The filing described Grok as one of four AI models capable of supporting national security applications and one of three equipped for top-secret mission-critical operations. Cameron Stanley, the Pentagon’s chief digital and artificial intelligence officer, wrote that the Grok Gov Model includes features found in no other frontier AI model and that restricting its use would severely impact the Pentagon.

Background on the lawsuit The NAACP filed the lawsuit in Mississippi, claiming xAI operates at least 57 turbines at its Colossus 2 data center without required pollution controls. The Department of Justice asked the court to dismiss the case, arguing the data centers provide critical energy capacity during armed conflict or other national security emergencies.

Stanley added that data centers powering government-used AI systems serve as a long-term strategic tool for maintaining technological advantage. The same filing noted that the National Geospatial-Intelligence Agency’s Maven Smart System uses AI to identify points of interest for military intelligence, though it does not create targets directly.

developments Several Democrats in Congress have proposed legislation that would require human commanders to retain control over life-and-death decisions and would prohibit AI use for nuclear weapons, domestic surveillance, and autonomous weapons systems.