In the first 24 hours of the assault on Iran, the US military struck more than 1,000 targets — nearly double the scale of the 'shock and awe' attack on Iraq two decades ago. This acceleration was driven by AI systems that revolutionized targeting, chief among them the Maven Smart System.
In her new book, Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare, journalist Katrina Manson investigates how Maven evolved from a 2017 experiment in computer vision for drone footage into a cornerstone of modern warfare. The project sparked employee protests at Google, the initial contractor, prompting the company to withdraw. A Marine intelligence officer named Drew Cukor pushed it forward, ultimately building the system with Palantir and technologies from Microsoft, Amazon, Anthropic, and others. Now used across US forces and recently purchased by NATO, Maven synthesizes satellite imagery, radar, social media, and dozens of other data sources to identify and target entities on the battlefield, dramatically speeding up the 'kill chain.'
Maven combines computer vision with a workflow management system that finds targets, pairs them with weapons, and allows users to click through the targeting cycle in seconds — a process that once took hours. An official told Manson that the technology enabled the US to go from hitting under a hundred targets a day to a thousand, and with the addition of LLMs, up to five thousand targets a day.
However, one of the thousand targets struck on the first day of the Iran war was a girls' school, killing over 150 people, mostly children. The school had previously been part of an Iranian naval base but was listed online as a school with visible playgrounds. Historian Kevin Baker argued in The Guardian that while some blamed AI hallucinations, 'People failed to update a database, and other people built a system fast enough to make that failure lethal.'
The pace of war is set to accelerate further. Manson reveals military programs developing fully autonomous weapons, including an explosive-laden drone Jet Ski capable of targeting and destroying targets independently.
I spoke to Manson about how AI is changing warfare.
This interview has been condensed and edited for clarity.
Colonel Cukor was an early and determined proponent of AI. Can you say a bit about him and his initial motivations?
He is the chief of Project Maven. He had a long-term vision stemming from frustration that US military operators in Afghanistan were equipped with poor intelligence tools. The US essentially fought that war 40 times over, every six months, because information wasn’t handed over when troops rotated. He wanted an analytic tool to bring intelligence to frontline operators, envisioning 'white dots' on a map infused with intelligence — coordinates, elevation, known details. This became a driving force behind Project Maven.
How was Maven initially conceived?
It started in 2017 as a project to use AI against satellite imagery, but was repurposed for drone video. The US wanted to develop AI for potential conflict with China, believing war would run faster than humans could think. Colonel Cukor proposed applying AI to drone video to analyze the vast footage; they sometimes managed to analyze as little as 4 percent of collection. AI would replace human eyes, but the vision was always bigger.
The public first heard about Maven during the Google protests in 2018. Google said the technology wouldn't be used to kill, but targeting was always the intention?
A Google spokesperson said flagging images for review was intended to save lives and was non-offensive. My reporting shows otherwise. US military operators aimed to save US lives and reduce civilian harm, so in that sense it's 'not offensive' as intelligence analysis. But in the wider sense, AI target selection was always intended for targeting. When I asked if targeting offensive weapon strikes were part of Project Maven, someone replied, 'Yeah, of course. The goal of the intel is to take out high-value targets.'
When Google backed out, Palantir stepped in. What was their role?
Microsoft and AWS took bigger roles in algorithms and compute. Cukor went to Palantir, pitching his 'white dots' vision and a user interface. Palantir initially didn't believe AI would take off and didn't want to just make a fancy interface — they wanted to crunch data. But Cukor was persuasive, and today Maven Smart System is becoming a 'program of record' with Palantir as prime contractor.
Ukraine sounded like a major inflection point. What happened?
The artillery fire team realized AI could speed up targeting, marking a crucial moment in the system's development.