OpenAI said it is updating its Pentagon contract after criticism over potential surveillance and military AI use.
CEO Sam Altman stated the company’s systems will not be used for domestic surveillance or by U.S. war intelligence agencies, citing constitutional and federal legal limits. Any use beyond those terms would require a separate contract change.
The clarification followed a deal to deploy OpenAI models on classified military networks, which sparked protests and internal dissent. Altman admitted the agreement was rushed and poorly communicated.
The controversy intensified after President Donald Trump halted federal use of Anthropic’s Claude system amid disputes over military AI safeguards. Nearly 500 OpenAI and Google employees backed Anthropic’s stricter limits, while advocacy groups organized protests against OpenAI’s Pentagon partnership.