OpenAI announced on Monday that it has amended its recent agreement with the Pentagon to explicitly prohibit the use of its models for domestic surveillance activities using commercially obtained data. This change follows criticism that the original agreement contained a loophole allowing such actions.

OpenAI CEO Sam Altman stated in a post on X that the new terms are "critical for protecting the civil liberties of American citizens." He added that due to significant public concern, the company wanted to make this point exceptionally clear.

OpenAI also noted that the Pentagon has confirmed that its services will not be used by intelligence agencies such as the National Security Agency (NSA) under the Department of Defense.

The agreement with the Pentagon was announced late last Friday. This development came after OpenAI's competitor, Anthropic, failed to secure a similar agreement because it refused to restrict the use of its AI models in certain scenarios, including domestic surveillance.

Although OpenAI had previously stated that, like Anthropic, it prohibits the use of AI for lethal autonomous weapons and large-scale domestic surveillance, legal experts pointed out that the original contract's phrase "all lawful purposes" left room for such activities. For example, existing legal loopholes allow government agencies to purchase sensitive personal data on the commercial market.