Despite OpenAI’s explicit policy prohibiting military and warfare applications of its technology, the US Department of Defense has tested OpenAI’s large language models. This testing was conducted through Microsoft, a major investor in OpenAI and a key partner that integrates its models into various products and services. The Pentagon’s Defense Innovation Unit (DIU) used Microsoft’s Azure OpenAI Service to explore the models’ potential for military tasks, such as querying open-source intelligence data and summarizing information. This arrangement highlights the complex relationship between AI developers, their commercial partners, and government agencies, raising questions about the enforceability of use-case restrictions when technology is licensed to third parties. The full article details this testing and the policy discussions it has prompted. Read the full article at: https://www.wired.com/story/openai-defense-department-ban-military-use-microsoft/