Pentagon Plans for AI Training on Classified Data

The Pentagon is exploring the establishment of secure environments where generative AI companies can train their models using classified data. This initiative aims to develop military-specific AI applications that could enhance operational capabilities. Currently, AI models, such as Anthropic’s Claude, are employed in classified settings primarily for answering queries and analyzing targets, including those in regions like Iran.

The proposed changes would enable these AI systems to learn from the classified data they encounter, potentially improving their effectiveness in military contexts. This move reflects a growing interest in leveraging advanced AI technologies for national defense purposes.

As the UK continues to navigate its own defense strategies, the implications of such developments in the US may influence discussions on AI in military applications within the UK government and defense sectors.

Source: www.technologyreview.com – https://www.technologyreview.com/2026/03/17/1134351/the-pentagon-is-planning-for-ai-companies-to-train-on-classified-data-defense-official-says/