OpenAI Amends Pentagon Deal to Specifically Bar Intelligence Agencies Like the NSA

In a surprising move following public criticism, OpenAI CEO Sam Altman has announced an amendment to the company’s recent Pentagon contract. The new “NSA Exclusion Clause” explicitly states that OpenAI technology will not be used by U.S. intelligence agencies, such as the National Security Agency, for surveillance or data mining activities. This addition was made to “make our principles very clear” after a weekend of intense public debate over the potential for government overreach.
The original deal between OpenAI and the Pentagon was struck just hours after the administration banned Anthropic. While the initial agreement already prohibited “domestic mass surveillance,” critics argued that the language was too vague to prevent intelligence agencies from using OpenAI tools to monitor foreign communications that might include American data. The new amendment is a direct response to those concerns, effectively putting the NSA on a “no-access” list for OpenAI services.
OpenAI’s decision to limit its own market is being viewed as a significant move to preserve its safety-first reputation. Altman admitted that the company “shouldn’t have rushed” the announcement on Friday, given the complexity of the issues involved. By adding these specific exclusions, OpenAI is attempting to build a wall between “lawful military operations” and “mass data surveillance,” a distinction that has historically been difficult to maintain.
The Department of War has reportedly accepted this amendment, reflecting a desire to maintain a stable partnership with OpenAI. This cooperation stands in stark contrast to the administration’s ” scorched-earth” policy toward Anthropic, which remains blacklisted and under legal threat. OpenAI is leveraging this stability to ensure its models are used for logistics, humanitarian aid, and strategic planning rather than covert spying.
As OpenAI moves forward with its classified deployment, it is also calling on the rest of the AI industry to adopt similar “transparency reports” for government work. OpenAI plans to periodically update the public on how its technology is being used by the military, within the limits of national security laws. This proactive approach is part of OpenAI’s strategy to remain a trusted global brand even as it becomes a critical defense contractor.