The removal of Anthropic’s AI system from the Pentagon has sparked a fight among other major artificial intelligence companies seeking to capitalize on this potentially lucrative opportunity to shape how AI is integrated into America’s military defense.
Earlier this month, the Pentagon called for Anthropic’s AI technology to be removed from military operations within six months. Intensifying feud between chief executive of the company And the Trump administration. internal pentagon memo He hinted that Anthropic’s artificial intelligence is being used in key areas of national security, including nuclear weapons, ballistic missile defense, and cyber warfare.
A source familiar with the U.S. military’s use of artificial intelligence told CBS News that the AI programs include one created by Anthropic, a company deemed by the Trump administration. Supply chain risk — Likely deployed as part of the U.S. operation against Iran.
The Pentagon hasn’t said exactly how the AI tools are being deployed, but CBS News spoke to several experts familiar with military operations to explain possible scenarios.
“The military currently processes approximately 1,000 potential targets a day and attacks the majority of them, and the time between attacks can be less than four hours,” said retired Navy Admiral Mark Montgomery, senior director of the Foundation for Defense of Democracies’ Center for Cyber and Technology Innovation. “Humans still know the information, but AI is now performing tasks that previously took days to analyze, and at a scale unmatched by previous campaigns.”
How to use AI in the military
The Department of Defense uses AI in the way many consumers do, summarizing and extracting a lot of information at once. By analyzing incoming documents, videos and images from the battlefield, AI can assist in military warfare scenarios to minimize casualties and determine which weapons are most effective, former Pentagon officials say.
“We are living through a military revolution driven by the digital revolution,” said CBS News national security analyst Aaron MacLean. “Today’s revolution is driven by an explosion of data: cameras everywhere, smartphones, connected cars, and more. The battlefield is now overflowing with information in ways that were unimaginable a generation ago.”
With so much data available, AI can help contextualize it for military personnel at a speed that far exceeds traditional human analysis.
“Currently, there is far more data than analysts can process on critical timelines. AI algorithms sift through it, build targeting packages, allocate strike assets, and assess damage, almost instantly,” MacLean said.
“The Israeli missile defense example shows this intuitively. When hundreds of drones and missiles come at you over a period of several hours, a human team cannot decide in real time which ones to intercept, what to use, and when to intercept them. That’s what AI does.”
To date, Anthropic’s large-scale language model, Claude, is the only large-scale AI system operating on top of the Department of Defense’s classified systems.
Josh Gruenbaum, head of the Federal Services Administration, the agency that helps governments decide which goods and services to use, said AI is also being used in other management functions such as research, policy development and procurement.
“Our goal is to help government agencies use this technology with confidence, significantly increasing output and efficiency for American taxpayers, while maintaining an unbiased approach that welcomes American innovators that strengthens the agency’s mission and ensures the government’s lawful adoption of these tools without undue obstacles,” Gruenbaum told CBS News.
How AI works with physical weapons
AI does not exist in isolation on the battlefield. There is still a lot of human surveillance and physical technology, including everything from aircraft carriers to drones to traditional defense contractors like Northrop Grumman, Boeing, and Lockheed Martin. The large language models that power the AI are used not to fly planes or fire missiles, but to do a lot of analysis before they are done.
According to Montgomery, this advancement has reduced work time from days to hours.
“This is a key enabler of the military’s ability to rapidly plan and execute war,” Montgomery told CBS News, stressing that while humans are still involved in the process, AI is being used to plan potential attacks.
A person with direct knowledge of Anthropic’s Claude AI’s military capabilities told CBS News that Claude’s primary job is to sift through large volumes of intelligence reports, including synthesizing patterns, summarizing findings, and surfacing relevant information faster than human analysts.
Officials said the targeting process remains human-driven. Anthropic’s U.S. Government Use Policy allows the Department of Defense to use its provisions to analyze foreign intelligence, but its terms of use require that all decisions regarding military targets be made by humans.
CBS News could not independently verify whether the Claude system was used in the incident. February 28th strike The incident hit an Iranian girls’ school, and the United States is likely involved.
AI greatly enhances operations, but wars can still occur without AI. Montgomery said the majority of weapons are still made by traditional contractors.
“This war is being fought with weapons, and 98 percent of it is being fought with conventional Prime-provided weapons, and they’re doing very well,” Montgomery said. He added that while it would be possible to wage war without AI, it would be “less desirable.” “It definitely plays a role in growing campaign after campaign, probably the next campaign,” he said.
Big Tech’s role in the military – and what’s changing
In July, the Department of Defense awarded a $200 million contract to artificial intelligence company Anthropic to integrate Claude into Department of Defense systems. The contract was subsequently terminated after a dispute arose between the Department of Defense and Antropic leadership over who should have final say in setting limits on how the military could use Claude.
now, The company is suing the federal government“The Constitution does not authorize the government to use vast powers to punish corporations for protected speech. There is no federal law authorizing the action taken here,” Anthropic said in its lawsuit.
Microsoft, OpenAI and Google employees have filed court briefs in support of Anthropic’s lawsuit.
The Department of Defense has instituted a six-month off-ramp period to remove Anthropic’s products from its systems, and Iran is still using the products despite the supply chain risk designation.
Meanwhile, other companies are also getting in on the action. Google announced in a blog post on Tuesday that it will deploy AI agents for unclassified military applications. Shortly after Anthropic’s clash with the Department of Defense in late February, Sam Altman, CEO of Anthropic rival OpenAI, posted on X about using the ChatGPT maker’s artificial intelligence models on the Pentagon’s classified networks. The company then posted about language in its contract with the Department of Defense that respects three red lines regarding the use of AI: lethal autonomous weapons, mass surveillance of American citizens, and high-stakes automated decision-making.
