Artificial intelligence is moving rapidly from research labs into the center of modern warfare, helping militaries analyze intelligence, use it for surveillance and accelerate battlefield decisions, as seen in the example of the current U.S.-Israeli war campaign against Iran, according to multiple media reports.
This shift, also reported during the war in Ukraine, could transform how wars are fought, analysts say, but it also raises serious legal and ethical concerns.
In the past weeks, the dispute between Claude-developer Anthropic and the U.S. Department of Defense (a.k.a Department of War) turned into an open example of why the anxiety voiced by some artificial intelligence critics and the urge for regulation and limitations to the development of powerful AI have been particularly emphasized since the launch of ChatGPT.
Late in February, U.S. Defense Secretary Pete Hegseth had given Anthropic's CEO Dario Amodei "a Friday deadline" to allow the company's artificial intelligence technology for unrestricted military use or risk losing its government contract. Amodei declined the call.
"We held to our exceptions for two reasons. First, we do not believe that today’s frontier AI models are reliable enough to be used in fully autonomous weapons. Allowing current models to be used in this way would endanger America’s warfighters and civilians. Second, we believe that mass domestic surveillance of Americans constitutes a violation of fundamental rights," the company said in a public statement following the U.S. administration's decision to designate it "a supply-chain risk."
Despite this, The Wall Street Journal (WSJ) reported that the U.S. Central Command (CENTCOM) used artificial intelligence tools developed by Anthropic in the operations in Iran, just "within hours" of declaring that the federal government will end its use of AI tools made by the very same company.
The same newspaper reported earlier that the U.S. military had allegedly used widely popular Anthropic's AI model Claude during the operation to capture Venezuelan President Nicolas Maduro at the start of the year.
The report, citing anonymous sources, at the time stated that Claude was used via Anthropic’s partnership with Palantir Technologies, a contractor for U.S. defense and federal law enforcement agencies.
As Amodei-led Anthropic has vowed to challenge the "supply-risk" designation in court, U.S. government agencies have shifted to alternative AI models, including OpenAI's ChatGPT, Reuters reported earlier this week.
On March 4, the company also said it received a letter from the Department of War confirming that "we have been designated as a supply chain risk to America’s national security."
In contrast, Anthropic's main rival, OpenAI, said it had struck a deal with the Pentagon to supply AI to classified U.S. military networks.
Announcing the deal, OpenAI CEO Sam Altman insisted that OpenAI’s agreement with the government included assurances that, for example, its systems would not be used for mass surveillance.
"Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems," he said on X. He added that the Pentagon "agrees with these principles, reflects them in law and policy, and we put them into our agreement."
However, only days after, Altman reportedly told employees that his company does not control how the Pentagon uses their AI products in military operations, The Guardian reported.
A different approach to close cooperation with the U.S. government on the sensitive issue of the use of technology for military purposes placed on different sides of the road the long-running rivals, Altman and Amodei, a former researcher at OpenAI. In 2021, Dario Amodei and his sister, Daniela, founded Anthropic along with other former senior members of OpenAI.
Amid this, a boycott campaign called "QuitGPT" has been spreading across the U.S. and beyond, with some celebrities such as Katy Perry openly putting their weight behind Claude and Anthropic.
Beyond Anhtropic-OpenAI cases, the debate over the use of AI, data analytics and surveillance in warfare was further demonstrated in the case of Iran.
On March 2, Iran, in retaliatory strikes, targeted an Amazon Web Services (AWS) data center in the United Arab Emirates (UAE). This was quickly tied to Israel's use of advanced technology and cloud services of leading American companies, including Google, Microsoft and Amazon.
The Israeli military is said to be storing large volumes of intelligence data collected on individuals in Gaza on AWS servers, according to exclusive reporting by 972 Magazine and The Guardian.
One dominant player in this regard is also Palantir, which, with platforms such as Maven Smart System (MSS), is providing services for the U.S. Department of Defense to identify, track, and target objects from satellite, drone, and sensor data. In 2024, the company also agreed to form a "strategic partnership" with the Israeli Defense Ministry to supply technology and support its war effort, just as Tel Aviv was expanding its offensive in Gaza.
Similarly, Israel is said to have had a comprehensive years-long intelligence campaign that helped pave the way for the recent assassination of Iran's Supreme Leader Ayatollah Ali Khamenei. The Financial Times (FT) report suggested that Israelis "were watching" the location where Khamenei was killed, pointing out that they had "nearly all traffic cameras in Tehran hacked for years."
Moreover, adding to the debate and controversy is that the use of artificial intelligence is not directly regulated in international humanitarian law, and that even defining what constitutes "fully autonomous weapons" may differ from directive to directive.
"For example, according to the International Committee of the Red Cross (ICRC), autonomous weapon systems are systems that select targets and use force without any human intervention. The U.S. Department of Defense, on the other hand, defines autonomous weapon systems as systems that, once activated, can select and strike targets without further intervention by the operator," Mustafa Tuncer, a faculty member at the National Defense University and an expert in international law, told Anadolu Agency (AA) recently.
"According to the U.S. approach, the operator of autonomous weapon systems may also have the authority to monitor the system’s activities and cancel them when necessary," he added.
Even more striking, claims were circulating online and on social media in recent days that AI might have been used to select an elementary school as a bombing target.
In the aftermath of airstrikes that leveled a school and claimed the lives of some 165 Iranian elementary students and staff, the Pentagon has refused to say whether the attack was suggested by an AI system, a report by Futurism publication said on March 6.
"We have nothing for you on this at this time," it quoted CENTCOM as saying.
