The numbers do not lie. In the first 24 hours of the assault on Iran, the US military struck more than 1,000 targets—an operational tempo that nearly doubled the legendary "shock and awe" campaign against Iraq over two decades ago. What made this possible was not a surge in personnel or aircraft. It was the Maven Smart System, an AI-powered targeting platform that has transformed from a 2017 research experiment into the engine driving the most concentrated application of artificial intelligence in warfare history.
The tension at the heart of this story is not whether AI makes war faster. It does. The tension is whether making war faster—with AI handling target identification, prioritization, and tracking—is a dangerous precedent that removes human judgment from the most consequential decisions a military can make.
According to a new book by journalist Katrina Manson tracing Maven's development, the system began as a modest experiment: applying computer vision algorithms to drone footage to help analysts distinguish between vehicles, buildings, and civilians. The initial contractor was Google, and when employees learned their work would support military targeting, they staged protests that made headlines worldwide in 2018. Google ultimately withdrew. The project continued anyway, passing through various contractors as it matured into operational capability.
Who wins in this scenario is clear: military planners who need to process battlefield data faster than any human team could manage. The Maven system can ingest hours of reconnaissance footage and satellite imagery, flag potential targets, and queue them for strike authorization. What commanders lose is harder to quantify—the friction of slower deliberation, the space for doubt, the human who might ask "but what if we're wrong about this building?"
The book, *Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare*, documents how the technology evolved from prototype to deployment. Military officials point to the speed and precision enabled by AI-assisted targeting. Critics see something else: a system optimized for throughput that may not adequately account for civilian presence, structural uncertainty, or the fog of war that no algorithm fully captures.
Those on the other side of this debate—military ethics scholars, human rights organizations, and technologists who study autonomous weapons—warn that once a threshold like 1,000 AI-assisted strikes in a single day becomes normalized, the expectation shifts. Faster targeting becomes expected. Slower deliberation becomes liability. The system that guides 1,000 strikes this week normalizes 2,000 next month.
What Maven's operational debut against Iran demonstrates is that AI warfare has crossed a threshold. This was not a limited test. This was not a controlled experiment. This was a full-scale military campaign where algorithms shaped which targets received ordnance and in what sequence. The human in the loop still exists—ostensibly, authorization remains with commanders. But when a system processes targeting data at machine speed and presents prioritized strike packages, the friction that slows human decision-making is engineered away.
The question now facing policymakers, military ethicists, and the defense technology industry is not whether this technology will be used. It is being used. The question is whether 1,000 strikes in 24 hours represents a new ceiling for AI in warfare—or merely a floor.