OpenAI is considering a potential contract to run its artificial intelligence technology on the North Atlantic Treaty Organization's unclassified networks, a person familiar with the discussions told Reuters on Tuesday. The consideration comes days after the company finalized an agreement to provide its systems to the Pentagon's classified environment.
Initial comments by the firm's chief executive on the scope of NATO deployment were later clarified. According to reporting, OpenAI's CEO at a company meeting indicated the company was seeking to deploy on all NATO classified networks. A company spokeswoman subsequently corrected that wording and said the NATO opportunity under consideration pertains to the alliance's unclassified networks.
NATO, a 32-member military alliance, did not immediately reply to requests for comment outside normal business hours.
OpenAI, which has financial backing from Microsoft and Amazon among others, announced a deal late last week to place its technology into the Pentagon's classified network. That move followed directives from U.S. President Donald Trump to cease work with Anthropic, a competitor that had been in contract talks with the Defense Department.
Anthropic was removed from consideration after a dispute during contract negotiations with the Pentagon over acceptable uses of the firm's models. Anthropic's chief executive, Dario Amodei, had emphasized the company would not allow its AI models to be used for mass domestic surveillance or to power fully autonomous weapons.
The Pentagon has previously stated it does not intend to employ AI to carry out mass surveillance of Americans or to develop weapons that function without human oversight. The Department has also expressed a desire to permit any lawful use of AI within its operations.
In an updated statement issued on Monday following the Friday announcement of the Pentagon agreement, OpenAI said its AI systems "shall not be intentionally used for domestic surveillance of U.S. persons and nationals." The statement added that the Pentagon affirmed AI services would not be used by intelligence agencies such as the National Security Agency.
According to reporting of an internal company meeting, OpenAI's CEO described the Pentagon arrangement as a complex decision with difficult branding consequences and negative public relations impacts in the short term. The CEO's remark was quoted as saying the choice was "an example of a complex, but right decision with extremely difficult brand consequences and very negative PR for us in the short term."
Context and implications
- OpenAI's interest in NATO's unclassified infrastructure follows its recent agreement with the U.S. Defense Department to deploy AI within classified networks.
- Clarifications from OpenAI indicate an intent to limit the NATO opportunity to unclassified systems, rather than classified networks as initially stated by company leadership.
- The episode underscores tensions in procurement discussions between defense customers and commercial AI firms, exemplified by the Pentagon's shift away from Anthropic amid disagreement over permitted uses of AI.