Pentagon's 'AI-first' push transforms military operations and intelligence analysis
The Pentagon is speeding up its use of artificial intelligence in military operations. A January directive from Defense Secretary Pete Hegseth has pushed the department to adopt an 'AI-first' approach. Generative AI tools are already assisting in combat, helping to create target lists and set strike priorities. The military plans to train AI models on intelligence data to improve their accuracy for specific tasks. Currently, human analysts handle these roles, but the Pentagon wants AI to take over. Before using classified data, officials will first test models trained on unclassified information to check their reliability.
The Pentagon has struck deals with companies like OpenAI and xAI to deploy their AI systems in secure, classified environments. Models such as Anthropic’s Claude are already being used to analyse targets, including those in Iran. However, it remains unclear which exact military tasks will require AI trained on sensitive data.
Training on classified material would happen in approved data centres. AI company employees with the right security clearances could access the information. Beyond combat roles, the military is also using AI for administrative work, such as drafting contracts and generating reports. The push for AI integration reflects a broader shift toward automation in defence operations. Models trained on classified data could eventually handle more complex analytical tasks. For now, the Pentagon is proceeding cautiously, evaluating performance before expanding access to sensitive information.
