Meta opens its AI models to US defence agencies and contractors

Meta Platforms Inc. has granted approval for US government agencies and defense contractors to use its AI models, opening the door for Meta's technology to play a key role in military and national security efforts.

The Facebook-parent company is making its large language models, called Llama, available to more than a dozen US agencies and contractors, including Lockheed Martin Corp., Booz Allen Hamilton Holding Corp. and Palantir Technologies Inc.

While Meta's AI models are open source, meaning they are free for developers to use, the company's "acceptable use policy" forbids people from using them on any projects related to "military, warfare, nuclear industries or applications, [and] espionage," among other things.

But Meta is making an exception for US defense agencies and their contractors as it works to expand its reach in the public sector and increase Llama's foothold in the burgeoning AI arms race.

"As an American company, and one that owes its success in no small part to the entrepreneurial spirit and democratic values the United States upholds, Meta wants to play its part to support the safety, security and economic prosperity of America - and of its closest allies too," wrote Nick Clegg, Meta's president of global affairs, in a blog post Monday.

Meta is also making Llama available to similar government agencies and contractors in the UK, Canada, Australia and New Zealand, a spokesperson confirmed. Those nations have an intelligence-sharing partnership known as Five Eyes.

Israel, the closest US ally in the Middle East, was not included on the current list of partners.

The move comes on the eve of the US election and at a time when Meta faces some concerns about the potential of developers in countries like China to use Llama for military applications.

Large language models are the foundation of the current AI boom and can be used for a variety of services, including powering chatbots. In the world of defense and security, Llama could help with tasks like data analysis and synthesizing documents, or to "track terrorist financing or strengthen our cyber defenses," Clegg wrote.

Meta will not have a say in how US agencies or its partners use the Llama technology, but a Meta spokesperson said via email that "it is the responsibility of countries leveraging AI for national security to deploy AI ethically, responsibly, and in accordance with relevant international law."

Meta's motivation to work with US defense agencies is driven in part by its intention to further establish Llama as a dominant foundation for AI products globally. Chief Executive Officer Mark Zuckerberg made Llama open source because he wants to have more control over the next major technological wave "- a strategy that only works if Meta's technology is ubiquitous.

On an earnings call last week, Zuckerberg said he was trying to get "the public sector to adopt Llama across the US government." The latest update could help with that goal and also pave the way for Zuckerberg to further develop relationships with elected officials.

Meta has been criticized over the years for its role in policing user speech and an alleged failure to protect young people online. Former President Donald Trump has taken several shots at Zuckerberg in recent months, even suggesting he might try and put the Meta founder in jail if he is re-elected to the White House.

Meta is also framing its decision to work with US defense agencies as a way of preventing US adversaries, like China, from gaining ground. "We believe it is in both America and the wider democratic world's interest for American open source models to excel and succeed over models from China and elsewhere," Clegg wrote.