Pentagon AI chief tells summit US is working with allies to avoid past mistakes

Maven Smart System helped strike 13,000 targets in Iran air campaign

Cameron Stanley, chief digital and artificial intelligence officer at the Department of War, told the Billington Cybersecurity summit on Tuesday that the United States is actively working with allied partners to keep them from repeating American missteps in adopting military artificial intelligence. Of NATO and the Five Eyes — the intelligence-sharing partnership of the US, UK, Canada, Australia and New Zealand — Stanley said Washington was “actively working with a number of our partners” so that “they don’t make the same mistakes that we make.”

On allied capacity, Stanley was blunt. “They don’t have the resources that we do, they don’t have the experience that we do, they don’t have the scale that we do,” he said, framing the gap as the reason US engagement matters.

The comments come less than a year into Stanley’s tenure, during which the Pentagon’s AI office has rolled out several new tools while losing a number of senior staff. Since January, Stanley has overseen the launch of GenAI.mil, an internal chatbot platform he says has reached 1.7 million users, and Agent Network, a system designed to shorten the time between intelligence-gathering and a commander’s decision. Pentagon secretary Pete Hegseth has directed Stanley’s office to report AI deployment speed as a monthly metric, part of a wider push to expand military computing power.

Much of Stanley’s appearance focused on speed. He described watching a commander halt a scheduled briefing to take a 30-second phone call and issue an order on the spot, without a formal decision brief or a review of the room’s security clearance level. He called the moment “revolutionary.”

“I know it sounds simplistic,” Stanley said. “It’s only simplistic because we’re good at it now.”

Project Maven, the Pentagon’s original AI targeting program launched in 2017, became the Palantir-built Maven Smart System, which the US leaned on heavily during its 2026 air campaign against Iran. According to Stanley, speaking at a conference in May, the system helped strike more than 13,000 targets over the first 38 days of the campaign. At Palantir’s AIPCon 9 conference in March, Stanley called Maven “revolutionary” too, saying it let commanders go from identifying a target to launching a strike within a single platform.

The campaign has drawn scrutiny. More than 120 members of the US Congress signed a letter demanding answers from the Pentagon regarding civilian casualties and the use of artificial intelligence, following a 28 February strike on an elementary school in Minab, which Iranian officials say killed at least 168 people, many of them children.

Stanley ended by telling the audience that the department’s old habits were now its biggest liability. “The way we’ve always done it is the most risky course of action,” he said, adding that the department needs to “think differently” if it wants AI tools in the hands of warfighters faster.