Academics studying the field say AI is significantly shortening the planning time for complex attacks, a phenomenon known as “decision compression.” Some warn that this could lead to a situation where military and legal experts simply formally stamp attack plans created by automated systems.
The use of AI-powered warfare tools to enable strikes on Iran is heralding a new era of faster-than-the-speed-of-thought bombing, experts have said, amid fears that human decision-makers could be overridden. Anthropic’s AI model, Claude, was reportedly used by the US military during the wave of attacks, as the technology “shortens the kill chain” – the process from target identification to legal approval to strike. The US and Israel, which had previously used AI to identify targets in Gaza, launched almost 900 strikes on Iranian targets in the first 12 hours alone. During those strikes, Israeli missiles reportedly killed Iran’s supreme leader, Ali Khamenei.
WAR WITH ARTIFICIAL INTELLIGENCE SHORTENS TIME
Academics studying the field say AI is significantly shortening the planning time for complex attacks, a phenomenon known as “decision compression.” Some warn that this could lead to a situation where military and legal experts simply formally stamp attack plans created by automated systems.
In 2024, Anthropic deployed its model to the U.S. Department of War and other national security agencies to accelerate military planning. Claude became part of a system developed by military technology company Palantir Technologies, in collaboration with the Pentagon, to “dramatically improve intelligence analysis and assist officials in decision-making processes.”
Craig Jones, senior lecturer in political geography at Newcastle University and an expert on “kill chains”, said: “The AI machine is making recommendations for targets, which in some ways is faster than the speed of thought itself. You have scale and speed at the same time: you are carrying out assassination-style strikes while paralysing the regime’s ability to respond with air-to-air ballistic missiles. This in historical wars might have taken days or weeks. Now it is all happening at once.”
HOW THE SYSTEM WORKS
The latest AI systems rapidly analyze vast amounts of information about potential targets: drone footage, telecommunications intercepts, and human intelligence. Palantir’s system uses machine learning to identify and prioritize targets, recommending the type of weapon, taking into account reserves and past performance against similar targets. It also uses automated reasoning to assess the legal basis for an attack.
David Leslie, professor of ethics, technology and society at Queen Mary University of London, called this “a new era of military strategy and technology.” But he warned of the danger of “cognitive overload”: the humans making the final decision could feel disconnected from the consequences because the analytical effort has been done by a machine. On Saturday, 165 people, many of them children, were killed in a missile attack on a school in southern Iran, according to state media. The facility was near a military barracks and the UN called the attack “a serious violation of humanitarian law.” The US military said it was investigating the reports.
GLOBAL COMPETITION FOR MILITARY ARTIFICIAL INTELLIGENCE
It is not known what AI systems, if any, Iran has integrated into its military machinery, although in 2025 it claimed to be using AI in missile guidance systems. Its AI program, hampered by international sanctions, seems negligible compared to superpowers like the US and China.
In the days before the attacks on Iran, the US administration had said it would exclude Anthropic from its systems after the company refused to allow the AI to be used for fully autonomous weapons or for surveillance of American citizens. However, the system remains in use until it is replaced. Meanwhile, its rival OpenAI signed a deal with the Pentagon for military use of its models. “The advantage lies in the speed of decision-making, in compressing planning from days or weeks to minutes or seconds,” Leslie said. “These systems produce options for human decision-makers, but they have a much narrower time window to evaluate recommendations.”
Prerana Joshi, a researcher at the Royal United Services Institute, added that the use of AI is expanding into all aspects of defence: logistics, training, decision management and maintenance. “AI is a technology that allows decision-makers to improve productivity and efficiency. It synthesizes data at a much faster pace, in order to aid the decision process,” she said. (The Guardian)