Pentagon says US military to be an 'AI-first' fighting force

Reuters

The US military plans to increase its use of artificial intelligence (AI) further after the Pentagon agreed to new and expanded contracts with some of the biggest names in technology.

Under eight agreements with Google, OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia and the start-up Reflection, the Pentagon said AI technology would now be used for any "lawful operational use".

"These agreements accelerate the transformation [of] the US military as an AI-first fighting force," the Pentagon said.

Conspicuous by its absence is Anthropic, as the company has said it is concerned about how the Pentagon could use its tools in warfare and domestically.

The firm is now suing the government over the alleged retaliation it faced after refusing to accept "any lawful use" language in its own contract.

The Pentagon on Friday noted that partnering with so many companies on AI would help it avoid "vendor lock" or being too reliant on a single company for its technology.

"Access to a diverse suite of AI capabilities from across the resilient American technology stack will give warfighters the tools they need to act with confidence and safeguard the nation against any threat," the Pentagon said.

It noted that more than a million people across the defence department had used the military's AI platform which hosts the tools since it was launched last year, helping them cut the time it took for many tasks "from months to days".

Access to powerful technology has become a key component of success in warfare and the Pentagon has been working to build up its AI capabilities for several years.

Anthropic's tools, including a version of its Claude chatbot, are still currently in use in many US government and defence agencies, as it was the first AI company to be deployed for classified work.

But earlier this year the relationship broke down as Anthropic chief executive Dario Amodei went public with fears that powerful AI tools could be used by defence agencies to conduct mass domestic surveillance and to deploy fully autonomous weapons of war.

Defence Secretary Pete Hegseth moved within days to label Anthropic a "supply chain risk" meaning it was deemed too dangerous for use in government settings.

Anthropic's legal challenge to that ruling is expected to go to court in September.

In the meantime the row appears to have opened the door for other AI companies to work more closely with the government and military.

OpenAI was the first company to ink a new deal with the Pentagon in the wake of Anthropic's row.

The ChatGPT-maker signed a contract at the end of February. A company spokeswoman said that Friday's announcement from the defence department was simply a formalisation of that deal.

"As we said when we first announced our agreement several months ago, we believe the people defending the United States should have the best tools in the world," the OpenAI spokeswoman said.

While Google's Gemini was also already in use by some parts of the government, this will be the first time the chatbot is being used to handle any government work at a classified level.

Earlier this week, hundreds of Google employees, including many from DeepMind, a part of the company that does much of the development work behind its AI models and tools, urged the company not to deepen its work with the government in a letter sent to chief executive Sundar Pichai and viewed by the BBC.

A Google spokesperson did not reply to a request for comment.

As for SpaceX, it is now the parent company of xAI, the AI startup Elon Musk formed after he acquired Twitter. The company operates the controversial AI chatbot Grok, but is widely considered to offer less advanced AI capabilities than the likes of Anthropic, OpenAI and Google.

A representative of SpaceX did not respond to a request for comment.

Nvidia and the startup Reflection will both have their open-source AI models, Nemotron and Reflection 70B, respectively, in use by the government. Nvidia is not providing any hardware as part of the deal. Nvidia declined to comment and Reflection did not respond to a request for comment.

Microsoft, AWS and Oracle have for years provided the cloud services purpose-built to enable government work that happens online. Microsoft and AWS did not reply to a request for comment, while Oracle said its defence work "enables the Department of War to build, deploy, and scale any model, without vendor lock-in."

Friday's announcement from the Pentagon marked a continuation of those services, which will now be used to deploy more AI models and tools than ever for military use.