OpenAI deploys ChatGPT to Pentagon platform as experts warn of security and human error risks

OpenAI Brings ChatGPT to Pentagon's Secure AI Platform

OpenAI announced it is deploying a custom version of ChatGPT on GenAI.mil, the Pentagon's secure AI platform developed by the U.S. Department of Defense 1

1

2

2

. This marks a significant expansion of US military access to advanced AI tools, as ChatGPT joins other generative AI models including Google Gemini and Grok from xAI on the Defense Department's platform 2

2

. The deployment follows months of internal deliberation at OpenAI and represents a major advancement for the OpenAI for Government platform launched in July 2025 1

1

.

Military Version Runs on Government Cloud Infrastructure

Unlike the consumer-facing interface, the military's ChatGPT will operate within authorized government cloud infrastructure with built-in safety controls and protections 3

3

. The system is approved specifically for unclassified work, allowing military personnel to use it for summarizing, analyzing, drafting documents, and generating reports 1

1

. OpenAI stated that safeguards are incorporated at the model level, though the Pentagon's ChatGPT will be available for "all lawful uses" without technical restrictions 1

1

3

3

.

OpenAI Wins Contract Over Competitors

OpenAI wasn't guaranteed the contract, as rival companies Google and xAI were also under consideration, both having accepted the "all lawful uses" clause 1

1

. Anthropic reportedly clashed with the Pentagon over restrictions, wanting more control over how Claude could be used, particularly regarding domestic surveillance and autonomous lethal operations 1

1

3

3

. OpenAI previously received a contract for up to $200 million with the Defense Department's Chief Digital and Artificial Intelligence Office last July as the agency accelerated AI adoption 3

3

.

Experts Warn of Security Risk from User Over-Reliance

Despite the built-in safeguards, critics warn that human error and user over-reliance on AI systems remain significant concerns for national defense applications. J.B. Branch, Big Tech Accountability Advocate at Public Citizen, cautioned that "when people use these large language models, they tend to give them the benefit of the doubt" 2

2

. This tendency becomes particularly problematic in high-impact military situations where accuracy is critical. Branch also warned that users often mistake AI tools for secure vaults, potentially exposing sensitive information to adversaries even when the system is designed for unclassified data only 2

2

.

Pentagon Accelerates AI Adoption Across Military Networks

The deployment comes as Defense Secretary Pete Hegseth announced plans in January to deploy leading AI models across both unclassified and classified military networks 2

2

. OpenAI emphasized that "we believe the people responsible for defending the country should have access to the best tools available" and that "AI can help protect people, deter adversaries, and prevent future conflict" 1

1

2

2

. While the initial rollout targets only the US military, OpenAI hinted at broader support for other democratic countries in the future, stating "our goal is to help governments use AI effectively and safely" 1

1

.