The company will permit selected authorities to deploy its AI for national security uses.

Meta is making its Llama artificial intelligence models available to authorities in France, Germany, Italy, Japan and South Korea, as well as European Union and NATO institutions, for defence-related tasks. Eligible bodies may also extend the waiver to contractors working on national security.

The change lifts Meta’s usual restrictions on military and security applications of its generative AI, introduced on Tuesday to bolster allied governments in Europe and Asia. The company said it acted in consultation with the United States government and would consider adding more countries after discussions with the Trump administration.

US military and national security agencies already use Llama-based tools, including a pilot that applies augmented reality to speed up repair work, indicating how the models could be applied in maintenance, logistics and training.

Meta is partnering with companies to deliver Llama-powered services, including Microsoft, defence technology firm Anduril and data analytics group Palantir. The move highlights the deepening role of major technology firms in defence modernisation among allied nations, while implementation choices will rest with national authorities and multilateral institutions.

Source: Euractiv.