Washington D.C. - OpenAI, the company developing artificial intelligence (AI), has signed an agreement with the U.S. Department of Defense (Pentagon) to use its AI tools in military systems. The company's Chief Executive Officer, Sam Altman, announced this information on Friday night via the social media platform X.
This decision was made public on the same day the administration led by President Donald Trump directed federal government agencies to stop using the AI tools of competitor company Anthropic.
Government agencies were told to stay away from the competitor's technology after Defense Secretary Pete Hegseth classified Anthropic as a risk in the supply chain. Generally, such a classification is used for companies with direct ties to foreign adversaries.
Anthropic, meanwhile, has maintained its stance that its AI should not be used for automated weapon systems or widespread surveillance of U.S. citizens. The company has also announced its intention to legally challenge the decision against it.
In the meantime, Altman claimed that the agreement with the Pentagon includes strict security and ethical standards. According to him, two key principles of OpenAI—a ban on widespread domestic surveillance and human responsibility in decisions regarding the use of force—are clearly stipulated in the agreement. He also stated that the Department of Defense has reflected these principles in its laws and policies.
Altman stated that OpenAI plans to develop technical safeguards to ensure its models behave as expected and to send engineers to the forefront at the Pentagon.
Furthermore, he mentioned that the Department of Defense has been urged to offer similar terms to all AI companies, expressing the view that a solution based on consensus should be pursued rather than legal disputes.
Pentagon Under Secretary for Technology, Emil Michael, also responded on X, stating that reliable and cooperative partners are necessary for military decisions involving life and death.
However, the actual differences in terms between OpenAI and Anthropic are not yet clear. Although further clarification has been sought from the Pentagon and OpenAI regarding this matter, detailed information has not yet been made public.
