On Monday, Meta announced a policy shift allowing its artificial intelligence models to be used for military purposes by U.S. government agencies and contractors involved in national security. Previously, Meta had restricted the use of its technology for such applications.

The company revealed that it would make its AI models, known as Llama, accessible to federal agencies. Additionally, Meta is collaborating with defense contractors such as Lockheed Martin and Booz Allen, as well as defense-focused tech firms like Palantir and Anduril. The Llama models are “open source,” enabling other developers, companies, and governments to freely copy and distribute the technology.

This marks a departure from Meta’s “acceptable use policy,” which previously barred the application of its AI software in military, warfare, and nuclear industries, among other fields.

In a blog post, Nick Clegg, Meta’s president of global affairs, stated that the company now supports “responsible and ethical uses” of its AI technology that align with the United States and “democratic values” in the global competition for AI leadership.

“Meta aims to contribute to the safety, security, and economic prosperity of America—and of its closest allies as well,” Clegg wrote, emphasizing that “broad adoption of American open-source AI models serves both economic and security interests.”

A spokesperson from Meta also confirmed that the company would share its AI technology with the Five Eyes intelligence alliance, which includes Canada, the UK, Australia, and New Zealand, in addition to the United States. Bloomberg initially reported Meta’s plan to extend its technology to the Five Eyes countries.

Meta, which owns Facebook, Instagram, and WhatsApp, has been actively working to make its AI software accessible to numerous third-party developers, competing with rivals like OpenAI, Microsoft, Google, and Anthropic in the AI sector. By open-sourcing its code, Meta hopes to bridge the gap with these companies. As of August, Meta’s AI software has been downloaded over 350 million times.

The company’s decision to allow military use may face criticism. Silicon Valley’s involvement in military projects has sparked controversy, with employees at Microsoft, Google, and Amazon voicing opposition to military-related deals.

Meta’s open-source approach to AI has also raised questions. While OpenAI and Google argue that releasing powerful AI tech openly could lead to misuse, Meta believes transparency can enhance safety by allowing widespread examination and improvement of its code.

Meta executives are reportedly concerned about potential regulatory pushback on open-source AI, especially following reports that Chinese government-linked research institutions had used Llama to develop software for the People’s Liberation Army. Meta disputed the report, asserting that the Chinese government does not have authorization to use Llama for military applications.

In his blog post, Clegg mentioned that U.S. agencies could employ Meta’s technology for tracking terrorist activity and strengthening cybersecurity across national systems. He emphasized that using Meta’s AI models could help the United States maintain a technological advantage over other countries.

“Our goal is to create a virtuous cycle, supporting U.S. technological leadership while making AI accessible worldwide and ensuring innovations are ethical, responsible, and align with the strategic interests of the United States and its allies,” he concluded.