Palantir defends AI targeting role as Maven becomes Pentagon staple
TL;DR: Palantir’s UK and Europe head Louis Mosley told the BBC that responsibility for how AI targeting platforms are used in war rests with military customers, not technology providers. The statement comes as the Pentagon formally designates Maven as a long-term programme and scrutiny grows over AI-assisted strikes during the US-Iran conflict.
The debate over AI in military decision-making has reached a new intensity. Palantir’s Maven Smart System — launched in 2017 to accelerate targeting by synthesising intelligence, satellite imagery, and drone footage — has reportedly been used to plan more than 11,000 US strikes against Iran since February.
Palantir’s position: a “support tool”
Speaking to the BBC, Mosley described Maven as a way for military personnel to “synthesise vast amounts of information that previously they would have had to do manually.” He rejected framing the platform as an automated targeting system, insisting that a human always makes the final decision.
But when pressed on the risk of commanders rubber-stamping AI recommendations under time pressure, Mosley redirected responsibility. “That’s really a question for our military customers. They’re the ones that decide the policy framework,” he said. This response — separating the tool’s design from its operational use — echoes a long-standing pattern among defence technology providers, and it is one that critics find insufficient.
Growing scrutiny from academics and lawmakers
Prof Elke Schwarz of Queen Mary University of London warned that the emphasis on speed “leaves very little time for meaningful verification of targets” and risks operators becoming over-reliant on AI outputs. US Democrats have also pushed back: Rep Sara Jacobs called for “strict guardrails” on military AI use, noting that “AI tools aren’t 100% reliable — they can fail in subtle ways.”
Questions have intensified following the deadly strike on a school in the Iranian town of Minab, where officials reported 168 casualties including around 110 children. Pentagon officials have faced questions about whether Maven contributed to target identification for that strike.
Looking forward
Despite the controversy, the Pentagon is doubling down. Last week, Reuters reported that Maven has been designated “an official programme of record” — meaning long-term integration across US military operations. With Anthropic’s Claude AI already phased out of the system over the company’s objections to autonomous weapons use, the question of which AI firms supply military platforms — and on what terms — will remain a defining issue for the industry.