Meta has announced an expansion of access to its Llama artificial intelligence models for national security purposes. The company stated that it is making the technology available not only to United States government agencies and their private sector partners, but also to a wider group of allied nations and institutions.
According to Meta, the Llama models are well-suited for sensitive applications because they are open source and can be securely downloaded and deployed without sending data through third-party AI providers. This allows governments to fine-tune the models with their own classified information, host them in secure environments, and use them on devices in the field.
Since late last year, Meta has provided Llama for national security uses to members of the Five Eyes alliance—Australia, Canada, New Zealand, and the United Kingdom—as well as their private sector collaborators. The company is now extending this access to additional democratic allies in Europe and Asia: France, Germany, Italy, Japan, South Korea, NATO, and European Union institutions.
Meta reported that Llama has been used by U.S. military and national security agencies to improve decision-making and operational efficiency. One example cited was a pilot project with the Army’s Combined Arms Support Command using AI alongside augmented and virtual reality technologies to speed up equipment repairs.
To support these efforts among U.S. allies, Meta is working with several industry partners including Accenture, Amazon Web Services, AMD, Anduril, Ask Sage, Booz Allen Hamilton, C3 AI, Circus Analytics & AI Research Lab (Circus), Cyberspatial Systems Inc., Databricks Inc., EdgeRunner AI Inc., Google Cloud Platform (GCP), IBM Corporation (IBM), Microsoft Corporation (Microsoft), Lockheed Martin Corporation (Lockheed Martin), Oracle Corporation (Oracle), Palantir Technologies Inc., Scale AI Inc., Snowflake Computing Inc., among others.
The company is also developing wearable products through a partnership with Anduril aimed at providing U.S. soldiers with enhanced perception and decision-making capabilities. Meta described this as "the largest effort of its kind" for equipping soldiers with advanced technology.
Meta emphasized that broad adoption of open source models like Llama will be important for maintaining U.S. leadership in artificial intelligence alongside its allies. The company referenced support for the U.S. government's AI Action Plan for America as part of this strategy.
Meta also noted that ethical deployment of AI in national security should follow international law and principles outlined in agreements such as the Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy.
"We are taking a step-by-step approach to extending access to Llama for defense and national security purposes," said Meta in its statement. "We will consider adding further countries in the future in consultation with the US government."