An artificial intelligence programme developed by the U.S. Department of Defense—known as “Project Maven”—is playing a central role in the ongoing military campaign against Iran, highlighting a profound shift in the conduct of modern warfare driven by rapid technological innovation, alongside complex ethical challenges.
What is Project Maven?
Project Maven is a pioneering AI initiative launched by the U.S. military in 2017, initially designed to assist analysts in processing the vast volumes of imagery captured by unmanned aerial systems. At the time, military analysts were required to manually examine each frame in detail to detect fleeting yet critical information. The programme was developed to enhance their ability to identify such details with greater speed and accuracy.
Eight years on, Maven has evolved significantly. It now supports targeting systems and battlefield management, dramatically accelerating the “kill chain”—the process of moving from detection to engagement.
How does it work?
Maven integrates air control and command functions, combining data from multiple sources, including sensors, satellite imagery, and intelligence on both friendly and adversarial forces. According to expert Alok Mehta, the system fuses these inputs to generate a comprehensive operational picture in real time.
The system analyses satellite imagery at high speed to detect movements and identify potential targets, effectively constructing an immediate overview of the operational theatre. This enables commanders to determine the most appropriate course of action for planned strikes.
During a recent Pentagon briefing, an official described Maven as operating “almost like magic,” transforming detected threats into actionable targeting options while evaluating possible responses and presenting commanders with a range of strategic choices.
Recent advances in generative AI have further expanded Maven’s capabilities, enabling users to interact with the system through natural language. This development has broadened access beyond technical specialists, making the technology more widely usable within military structures.
This capability has been supported by the “Claude” model developed by Anthropic. However, this partnership may not endure, as the Pentagon reportedly penalised the company after it publicly refused to allow its AI tools to be used for fully autonomous strikes or domestic surveillance.
Why did Google withdraw?
From its early stages, the use of AI in military contexts has raised significant ethical concerns. Google was the original technology provider for Project Maven, but in 2018, more than 3,000 employees signed an open letter opposing the company’s involvement, arguing that the project crossed ethical boundaries. Several engineers resigned in protest.
Subsequently, Google withdrew from the contract, which had aimed to improve drone capabilities in distinguishing between objects and human targets using AI. The company later published a set of AI principles explicitly ruling out participation in weapons systems.
This episode underscored a broader divide between engineers in Silicon Valley—many of whom view autonomous targeting as an ethical red line—and military officials, who increasingly regard such technologies as operationally unavoidable.
Over time, however, Google has eased some of these restrictions and signalled a renewed willingness to engage in defence-related contracts. The Pentagon has indicated that Google, alongside OpenAI and Elon Musk’s xAI, is now competing to replace Anthropic’s “Claude” within Project Maven.
The role of Palantir
In 2024, Palantir—a company founded with partial funding from the U.S. Central Intelligence Agency and originally focused on intelligence applications—stepped in to fill the gap left by Google. It has since become the primary supplier for Project Maven, with its AI technologies forming the programme’s operational backbone.
Palantir CEO Alex Karp has argued that the world is increasingly divided between those who possess such technologies and those who do not.
What are the results so far?
Both the Pentagon and Palantir have declined to comment on Maven’s specific performance in the Iran campaign. However, the pace of U.S. strikes suggests that the system has significantly accelerated targeting and engagement processes.
Within the first 24 hours of the joint operation with Israel, U.S. forces reportedly conducted strikes on more than 1,000 targets—an indication of the transformative impact AI systems like Maven are having on modern warfare.
