This post is also available in:
עברית (Hebrew)
As the volume of information flowing into military headquarters continues to grow, decision-making becomes increasingly complex. Satellites, drones, reconnaissance aircraft, intelligence systems, and other sources generate vast amounts of real-time data. For modern militaries, the challenge is no longer simply collecting information, but rapidly analyzing it, identifying targets, and prioritizing actions within extremely compressed timeframes.
Against this backdrop, documents submitted to a U.S. court last week reveal that the Pentagon relied on a specialized version of xAI’s Grok artificial intelligence model during operations in Iran. According to testimony provided by a senior U.S. Department of Defense official, the system supported operations in which more than 2,000 munitions were launched against thousands of targets over a period of 96-hours.
The disclosure offers a rare glimpse into how AI systems are being integrated into modern targeting cycles. Contrary to the popular image of a “robot deciding when to shoot”, these systems are generally used for information processing and analysis: identifying potential targets, correlating intelligence sources, prioritizing threats, and presenting recommendations to human operators. As data volumes continue to increase, AI’s ability to filter, organize, and prioritize information is becoming a critical component of military operations.
According to the documents, the Pentagon is using a dedicated version known as Grok Gov, which was developed specifically for government agencies and designed to operate within classified environments. The platform joins a growing list of AI tools already integrated into defense programs, including systems for geospatial intelligence analysis, object recognition in aerial imagery, and real-time intelligence processing.
At the same time, the use of artificial intelligence in military systems remains controversial. In the United States, lawmakers are advancing legislative proposals aimed at ensuring that life-and-death decisions remain under human control. The debate centers on where the line should be drawn between systems that provide recommendations and those that directly influence target selection and the use of force.
From a defense perspective, the case highlights a broader trend: a gradual shift from military capabilities centered primarily on weapons and platforms toward systems in which information processing plays a decisive role. In an era where the pace of warfare is measured in seconds, the ability to understand massive volumes of data in real time may prove just as important as the ability to launch the munition itself.
