The United States government has acknowledged for the first time that Elon Musk’s Grok artificial intelligence was deployed in military operations against Iran, a disclosure buried in a legal motion to dismiss an environmental lawsuit. The Department of Justice argued that shutting down the gas turbines powering xAI’s massive data centre in Memphis would imperil national security, citing sworn testimony from the Pentagon’s AI chief, Cameron Stanley. He confirmed that a derivative of Grok, the Grok Gov Model, was integrated into Project Maven, the algorithmic warfare programme that identified and prioritised targets during a 96-hour operation known as Epic Fury, during which more than 2,000 munitions were launched. The revelation, first reported in a Justice Department brief dated 15 June, has abruptly dragged a commercial chatbot into the heart of the most sensitive debate over autonomous warfare.

Viewed from Washington, the administration’s intervention is a muscular defence of Musk’s business empire. The NAACP and local community groups filed suit alleging that xAI’s Memphis facility operates dozens of natural gas turbines without proper permits, spewing pollutants near homes, schools and churches in predominantly Black neighbourhoods. The Justice Department’s motion insists that the state of Mississippi, not federal courts, should adjudicate the dispute, and that any disruption to the data centre’s power supply would threaten “national, economic and energy security.” The filing effectively weaponises a national security argument to shield a private company from civil rights litigation, a move that critics say sets a troubling precedent for corporate impunity.

From Tehran, the admission has landed with explosive force. Iranian media have seized on the Pentagon’s statement, linking Grok’s targeting algorithms to a strike on a primary school in Minab that killed 168 children. While the US has not confirmed that specific incident, the broader acknowledgment that AI was used to select targets in densely populated areas has intensified accusations of unlawful killing. European analysts note that the disclosure exposes the extent to which commercial AI tools, developed with little public oversight, are now embedded in the Pentagon’s kill chain. The Italian and German press have framed the story as a cautionary tale about the dual-use nature of generative AI, with Musk’s xAI simultaneously powering consumer chatbots and military targeting systems.

The convergence of environmental justice, corporate power and automated warfare in a single court filing marks a new chapter in the governance of artificial intelligence. The Pentagon’s reliance on xAI’s infrastructure means that a pollution dispute in Tennessee now carries implications for US military readiness and the legality of strikes under international humanitarian law. Analysts in London note that the case exposes a regulatory vacuum: no clear framework exists to assess whether AI models trained for civilian use can be responsibly repurposed for lethal operations. As Musk oscillates between roles as tech entrepreneur, government efficiency adviser and diplomatic envoy, the blurred boundaries between his commercial ventures and state interests are likely to face intensifying scrutiny from Congress, allied governments and human rights bodies.