Pentagon Says It Used Elon Musk’s Grok to Help Identify Targets in Strikes Against Iran

The Pentagon has disclosed that U.S. military personnel used Elon Musk’s artificial intelligence tool, Grok, to help identify targets for strikes against Iran, according to a senior defense official speaking at a briefing on Tuesday. The disclosure adds to a growing debate in Washington over how commercial AI systems are used in national security and military operations.

The official did not describe in detail what specific steps in the targeting workflow relied on Grok, nor did the Pentagon publicly provide a breakdown of how outputs were verified before action. The disclosure, however, indicates that the system was used for target identification purposes rather than purely for general analysis or administration.

Grok is a commercial AI chatbot offered by the companies associated with Elon Musk. The Pentagon’s acknowledgment raises new questions about the role of private-sector AI tools in sensitive defense functions, particularly where errors could lead to harm to civilians or unintended escalation.

U.S. officials have previously emphasized that military targeting processes involve multiple layers of review and compliance requirements. In this case, the Pentagon’s statement focused on the fact that Grok was used in identifying targets, without providing additional information about safeguards, documentation, or the duration of its use.

The revelation comes amid broader scrutiny of AI systems and their governance, including questions about transparency, data handling, and accountability when commercial tools are connected to defense workflows. Lawmakers and watchdog groups in the United States have repeatedly pressed agencies to clarify what tools are used, how they are evaluated, and who is responsible when outcomes are contested.

No additional details were provided in the report about the specific strike episodes, geographic locations, or the date range during which Grok was used. The Pentagon’s disclosure therefore leaves key parts of the operational and oversight record unclear from public information.

The Pentagon did not describe an immediate change in policy in response to the disclosure, but the acknowledgment is likely to draw follow-up questions from Congress and oversight bodies about procurement, approval, and the testing and validation of AI systems for operational use. Any further public clarification would shape how agencies explain compliance with existing targeting authorities and rules designed to protect civilians.