As the U.S. military expands its use of AI tools to pinpoint the targets of Iranian airstrikes, lawmakers are calling for guardrails and increased oversight of the technology’s use in warfare.
Two people familiar with the matter, who requested anonymity to discuss sensitive matters, confirmed that the military is using AI systems from data analytics firm Palantir to identify potential targets for ongoing attacks. The use of Palantir’s software, which relies in part on Anthropic’s Claude AI system, comes as Secretary of Defense Pete Hegseth seeks to put artificial intelligence at the center of U.S. combat operations and clashes with Anthropic’s leadership over limits on the use of AI.
But as AI takes on a broader role on the battlefield, lawmakers are calling for greater focus on safeguards that govern its use and greater transparency about how much control is transferred to the technology.
Rep. Jill Tokuda (D-Hawaii), a member of the House Armed Services Committee, told NBC News in response to questions about the use and reliability of AI in the military that “a full and impartial review is needed to determine whether AI is already harming or threatening human life in the war with Iran.” “Human judgment must remain at the center of life-and-death decisions.”
The Department of Defense and major AI companies such as OpenAI and Anthropic have publicly stated that current AI systems should not be able to kill without human approval. However, concerns remain that relying on AI for some operations and decision-making could lead to errors in military operations.
Pentagon chief spokesman Sean Parnell told X on February 26 that the military “does not want to use AI to develop autonomous weapons that operate without human involvement.”
The Pentagon did not respond to questions about how the military balances the use of AI to reduce human workload while verifying that analysis and targeting recommendations are accurate.
Lawmakers and independent experts who spoke to NBC News warned against the military’s use of these tools and called for clear safeguards to ensure humans remain involved in life-and-death decisions on the battlefield.
“AI tools are never 100% reliable, and operators continue to place too much faith in them, even though they can fail in subtle ways,” said Rep. Sarah Jacobs (D-Calif.), a member of the House Armed Services Committee.
“We have a responsibility to put strict guardrails against the military’s use of AI and ensure that humans are involved in every decision to use lethal force, because the wrong decisions can be catastrophic for the civilian and military personnel performing these missions,” she said.
Anthropic’s Claude is a key component of Palantir’s Maven intelligence analysis program, which was also used in the U.S. operation to capture Venezuelan President Nicolas Maduro. A person familiar with Anthropic’s work with the Pentagon said Claude is used to help military analysts sift through information and not directly provide targeting advice.
News about Claude’s role in recent military operations was first reported by the Wall Street Journal and the Washington Post.
But that role was complicated by a clash between Anthropic and Hegseth after Anthropic sought to prevent the military from using AI for domestic surveillance and autonomous lethal weapons. The Pentagon last week recognized Anthropic as a threat to national security and said it may remove it from military use in the coming months. Anthropic filed suit to challenge this designation.
Antropic declined to comment. Palantir did not respond to a request for comment.
Adm. Brad Cooper, the commander of U.S. Central Command, acknowledged in a video posted to X on Wednesday that AI has become a key tool in helping the U.S. select targets in Iran.
“Our warfighters utilize a variety of advanced AI tools. These systems help sift through vast amounts of data in seconds, allowing our leaders to cut through the noise and make smart decisions faster than our adversaries can react.”
“While humans will always make the final decisions about what to photograph, what not to photograph, and when to photograph it, advanced AI tools can reduce processes that previously took hours or even days to seconds.”
The Trump administration has publicly embraced the use of this technology both in the military and throughout the government.
Rep. Pat Harrigan (R.N.C.) said AI has already become essential to quickly processing military intelligence, including on Iran.
“AI is a tool that helps warfighters process vast amounts of data faster than humans alone could,” Harrigan, who also serves on the House Armed Services Committee, told NBC News in a statement.
“But no AI system can replace the judgment, training, and experience of America’s warfighters. Having a human in the loop is not a formality, it is a necessity, and there is nothing in the way our military operates to suggest otherwise.”
Lawmakers interviewed by NBC News did not say AI should be completely excluded from military use, but some said more oversight is needed.
Sen. Elissa Slotkin (D-Mich.), a member of the Senate Armed Services Committee, said the Pentagon is not doing enough to clarify the extent to which humans vet AI-assisted or generated military information.
“It’s really up to people, in this case the Secretary of Defense, to ensure that we have a surplus of personnel for the foreseeable future, and we have no confidence whatsoever in that,” she said.
Sen. Mark Warner (D-Va.), the top Democrat on the Senate Intelligence Committee, said he is concerned about the military’s use of AI to help identify targets and said there are unanswered questions about how the new technology is being used. “This has to be addressed,” he told NBC News.
OpenAI and Anthropic, which have worked with the U.S. military, say even their most advanced systems are error-prone, and the world’s top AI researchers admit they don’t fully understand how the most advanced AI systems work.
“Even with the systems we’ve built, we can’t say there’s a 100% chance that they’ll be completely reliable,” Anthropic CEO Dario Amodei said in an interview with NBC last month.
A major OpenAI study published in September found that all major AI chatbots, which rely on systems called large-scale language models, “hallucinate” or regularly fabricate answers.
Sen. Kirsten Gillibrand (D.Y.) called for clearer rules on how the military can use AI.
“The Trump administration has already demonstrated that it is willing to subvert American law to wage an unpopular war,” she told NBC News. “Absent explicit safeguards, there is little reason to believe that the Department of Defense will have any further liability for the use of AI.”
Mark Beal, director of government affairs at the AI Policy Network think tank in Washington, D.C., and director of AI strategy and policy at the Pentagon from 2018 to 2020, said that while AI can streamline the process of deciding where to attack, it’s clear that humans still need to thoroughly vet targets.
“There are many steps before the trigger is pulled. AI systems have been deployed very effectively to accelerate existing workflows, giving commanders, analysts and planners better and faster decision-making capabilities,” he added. “However, when it comes to actually fielding weapon systems, this technology is not yet ready.”
“These systems are going to be very good, and as other adversaries start using them, there will be even more pressure to shorten reviews of AI output to get it working at useful and effective speeds,” Beer said. “Before we get there, we have to find a way to solve this reliability problem. No matter what you think about lethal autonomous weapons, it’s in the whole world’s interest to make them safe and effective.”
Heidi Klaaf, chief scientist at the AI Now Institute, a nonprofit that advocates for the ethical use of the technology, said she worries that relying on AI to quickly process life-or-death information could be a way for the military to avoid responsibility for mistakes.
“Given how inaccurate these models are, it is extremely dangerous that we are somehow being sold that ‘speed’ is strategic here, when in reality it is a cover for indiscriminate targeting,” Klaf said.
