The Iran war is adding new urgency to a widening global debate over who controls the future of AI as a tool of war, including whether the rapidly evolving technology can be used in a lawful manner.
It lies at the heart of a high-stakes dispute pitting US defence officials against Anthropic PBC, one of the most promising AI companies whose models are used on the Pentagon’s classified cloud.
After failing to agree with Anthropic on terms governing the use of its AI technology, US Defence Secretary Pete Hegseth declared the company a supply chain risk and gave military contractors six months to stop working with the firm.
US President Donald Trump also instructed federal agencies to cease work with Anthropic, describing it as an “out-of-control, Radical Left AI company”.
Anthropic Chief Executive Officer Dario Amodei and defence officials have since resumed discussions after last week’s breakdown in talks, raising the possibility that the Pentagon could reach an agreement with the company and avert the penalties threatened by Hegseth.
Although Amodei has expressed alarm at using AI in fully autonomous weapons before the technology is reliable and does not want his firm’s tools used to spy on US citizens en masse, he does support working on lethal US military operations that abide by those red lines.
Among the AI tech used in the Iran campaign is Maven Smart System, a digital mission control platform, according to people familiar with the US operations, who spoke on condition of anonymity in order to discuss sensitive information.
The system, produced by Palantir Technologies Inc, is fed by more than 150 different data sources, according to previous public statements from US military officials.
That system is also using Anthropic’s Claude AI tool among the large language models installed on the system, according to the people, who said Claude is working well and has become central to US operations against Iran and to accelerating Maven’s AI efforts.
Spokespeople for Anthropic declined to comment while representatives for Palantir didn’t immediately respond to requests for comment. The Washington Post previously reported the US military’s use of Maven Smart System and Claude in US military operations.
Anthropic isn’t the only AI-enabled defence tech company involved in the Iran war.
Trae Stephens, co-founder and executive chairman of Anduril Industries Inc, told Bloomberg TV that his company’s technology was being used in the conflict.
“We have all sorts of primarily counter-air systems that are present in conflict zones,” he said, adding that the company was “actively working day to day” with the Defence Department on ongoing operations.
“Obviously I can’t give a whole lot of details beyond that,” he said. The defence technology company has said it aims to remake the military of the US and its allies with AI, speedy manufacturing and other technology.
Hawkins, the Central Command spokesperson, said that artificial intelligence helps analysts whittle down what they need to focus on, generating so-called points of interest and helping personnel make “smart” decisions in the Iran operations. AI is also helping to pull data within systems and organise information to provide clarity, he said.
“Bottom line, these tools help leaders – humans – make smarter decisions faster. The tools do not replace them or make targeting decisions,” said Hawkins, adding that target selection relies on a very specific, rigorous, legal process that involves commanders and leaders.
Some organisations, such as Stop Killer Robots, a coalition of 270 human-rights groups, argue that AI-enabled decision-support systems reduce the separation between recommending and executing a strike to a “dangerously thin” line and risk introducing automation bias, when humans overly trust outputs produced by machines.
Centcom is investigating possible incidents of civilian harm following reports that a strike against a girls’ primary school killed more than 160 people. There’s no indication whether AI played a role.
“We take these reports seriously and are looking into them,” Hawkins said. “The protection of civilians is of utmost importance, and we will continue to take all precautions available to minimize the risk of unintended harm.
“Unlike the Iranian regime, we have never – and will never – target civilians.”
– With assistance from Shirin Ghaffary and Ed Ludlow.
Sign up to Herald Premium Editor’s Picks, delivered straight to your inbox every Friday. Editor-in-Chief Murray Kirkness picks the week’s best features, interviews and investigations. Sign up for Herald Premium here.
