Written by Tiffine Le Riboud and Thomas Urban
Paris, France (AFP) March 7, 2026
Recent battles between the United States, Israel, and Iran have deployed AI like never before to sift through information and select targets, but the use of this technology in war remains hotly debated.
Various forms of artificial intelligence are reportedly being used in the Israeli operation in Gaza and in the American raid that captured Venezuelan leader Nicolás Maduro.
And while experts believe the technology has helped select targets for the thousands of U.S. and Israeli attacks on Iran since February 28, its exact use has yet to be confirmed.
Laure de Lucie Rochegonde, of the French think tank IFRI, said that currently “all important military powers are investing heavily in military applications of AI.”
“Almost every military function, from logistics to reconnaissance, observation, information warfare, electronic warfare, and cybersecurity, can be enhanced with AI,” he added.
We can see AI tools being incorporated into semi-autonomous attack drones and other weapons as well.
But one of its best-known uses is to reduce the so-called “kill chain,” or the time and decision-making between detecting a target and attacking it.
The U.S. military uses the Maven Smart System (MSS) built by Palantir, which the company says can identify and prioritize potential targets.
The Washington Post reported this week that Anthropic’s Claude-generated AI model has been integrated with Maven to enhance the tool’s detection and simulation capabilities.
Palantir and Anthropic did not respond to AFP’s requests for comment.
Bertrand Lombier, head of the French army’s AI agency AMIAD, said AI algorithms “allow us to process information much faster and, above all, more comprehensively.”
The technology can also sift through vast amounts of data, including “satellite imagery, radar, electromagnetic waves, sound, drone imagery, and in some cases real-time video,” he added.
– Human control –
Deploying AI in warfare raises a number of moral and legal questions, particularly regarding the extent of human control over AI behavior.
This argument surfaced during the fighting in Gaza, where the Israeli military used a program called “Lavender” to identify targets within a certain margin of error.
The application worked “because it covered a very limited area,” de Lucie Rochegonde said.
Israel also has a “mass surveillance system” that can feed Lavender with data about the enclave’s inhabitants.
“It seems unlikely that such a system is being built in Iran,” she added.
“If something goes wrong, who will take responsibility?” Peter Asaro, chairman of the International Commission on Robotics and Arms Control (ICRAC), told AFP in an interview.
He added that the widely reported school bombing in Iran, in which Iranian authorities say 150 people were killed, may have been the result of mistaken targeting by AI.
Neither the United States nor Israel have acknowledged responsibility for the attack.
AFP was unable to visit the school and was unable to confirm what happened there.
However, the site was near two facilities controlled by Tehran’s powerful ideological elite, the Islamic Revolutionary Guards Corps (IRGC).
“They didn’t distinguish it from the military base it should be. (But) who are they?” he asked – humans or machines?
If AI is used, he argued, the key questions are “how old is the data” used for targeting and whether the erroneous attacks were due to “database errors.”
-Step by step-
Lombier said it was “science fiction” for AI to “operate without anyone controlling it.”
At least in France, he argued, “military commanders are at the center of the actions and the design of these systems.”
“Military decision-makers will not consent to the use of AI unless they have trust and control over the behavior of the AI,” Lombier added.
“They know what the risks are, what the capabilities of these systems are, and in what situations and with what level of confidence they can be used.”
Benjamin Jensen of the Washington-based think tank CSIS, which has participated in testing AI in military decision-making for the past decade, said today is just the “beginning” of AI use by the world’s militaries.
The world’s militaries have not “fundamentally reimagined how they plan and conduct operations” with AI capabilities, he added.
“It’s going to take a generation to really understand this.”
tu-tll/tgb/ah/pdw/cms
Related links
All about robots on Earth and beyond!
