OpenAI announced on 9 February 2026 that it is bringing ChatGPT to GenAI.mil, a U.S. government domain for generative AI services [source]. The deployment marks the first time ChatGPT will be available on a .mil domain, indicating integration with U.S. military or Department of Defense infrastructure.

The announcement provided limited technical detail about the deployment. OpenAI stated that the service would be available to authorized users on the GenAI.mil platform, but did not specify which model version would be deployed, whether the system would operate on government-controlled infrastructure, or what data handling protocols would apply.

No information was provided about training data restrictions, output filtering, or whether the deployment would use a distinct model variant from the public ChatGPT service. The announcement did not address whether user interactions would be logged, retained, or used for model improvement.

The GenAI.mil domain appears to be a centralized platform for AI tools within the U.S. defense ecosystem, though OpenAI's post did not clarify the platform's governance structure or operational oversight.

This deployment follows a pattern of AI providers entering government contracts without public disclosure of model behaviour differences or safety constraints specific to those environments. Previous government AI deployments have raised questions about transparency when the same models serve both public and classified use cases.

OpenAI did not respond to questions about whether this deployment involved modifications to ChatGPT's base capabilities or content policies. The announcement contained no information about independent auditing, red-teaming, or third-party verification of the system's performance in a military context.