OpenAI Adds Custom ChatGPT to Pentagon Platform as Experts Warn of Risks

OpenAI has deployed a customized version of its ChatGPT model to the U.S. Department of Defense’s GenAI.mil platform. The system is designed to support unclassified Defense Department operations. The move expands the military’s access to generative AI tools amid growing interest in their capabilities.

The ChatGPT integration comes alongside other AI models already available on GenAI.mil, including Google’s Gemini and xAI’s Grok according to reports. OpenAI emphasized that the platform includes safeguards to protect sensitive data and ensure it is not used to train public models.

Experts caution that reliance on AI systems can introduce risks, especially when users place too much trust in their outputs. J.B. Branch of Public Citizen noted that users may mistakenly assume AI tools are secure vaults for sensitive data.

The Pentagon launched GenAI.mil in December 2025 as part of a broader push to integrate advanced AI capabilities. The platform now has over one million unique users.

Pentagon officials, including Defense Secretary Pete Hegseth, have actively encouraged adoption of the platform to improve operational tempo and decision-making.

OpenAI’s involvement follows a $200 million contract with the Pentagon in June 2025. Similar contracts were awarded to Google, xAI, and Anthropic for their AI models.

How Did Markets React?

While no direct market impact was reported, the announcement highlights the growing importance of AI in enterprise and government settings. Companies like Oracle and Anthropic are already seeing shifts in investor sentiment due to evolving AI competition.

OpenAI’s recent focus on enterprise AI solutions, such as the Frontier platform, indicates a broader strategy to compete with traditional software providers like Palantir and C3.ai.

What Are Analysts Watching Next?

Analysts are closely monitoring the integration of AI models into sensitive environments. J.B. Branch warned that even unclassified data can pose risks if not properly managed.

The Pentagon is also looking to expand its GenAI.mil platform in the coming months, potentially adding a fourth AI model. Anthropic’s Claude has yet to be integrated, despite recent security concerns.

Investors should watch for updates on how the military balances AI adoption with data security, especially as more powerful models become available. OpenAI’s continued focus on enterprise AI solutions could also shift the competitive landscape in the tech sector.

OpenAI did not immediately respond to a request for comment by Decrypt.