Altman Says OpenAI Defense Contract Includes Surveillance Safeguards

[ed]

Despite the political fallout surrounding Anthropic, Sam Altman says OpenAI’s agreement with the United States Department of Defense includes explicit protections addressing the same ethical flashpoints. In a statement posted to X, Altman emphasized two core principles: prohibitions on domestic mass surveillance and maintaining human responsibility in the use of force, including autonomous weapons systems. He added that these standards are reflected in existing US law and embedded directly into the contract.

Altman also revealed that OpenAI will build a dedicated “safety stack” to prevent misuse of its models and deploy engineers alongside Pentagon teams to ensure compliance. According to reporting from Fortune, Altman told employees that if a model refuses a task, the government will not compel OpenAI to override that refusal. As geopolitical tensions escalate — including reports of US and Israeli military action against Iran — the agreement places OpenAI at the center of a rapidly evolving intersection between AI development, national security, and public trust.

[mendez]

“human responsibility in use of force” meanwhile the model just wanted to write emails