OpenAI enters into a contract with the Pentagon to develop military AI applications. According to OpenAI, the intention is not to turn AI into a weapon of war.
The US Department of Defense has placed OpenAI on its list of approved contracts. OpenAI receives a check worth 200 million dollars for developing “groundbreaking AI capabilities”. An initial advance of 2 million dollars will be paid immediately, with the remainder to be received by OpenAI after the contract expires in July 2026.
In a blog post, OpenAI writes that the pilot project with the US military is just the start of a broader collaboration with the government, under the newly announced initiative OpenAI for Government. The collaboration means that OpenAI will provide government services with secure access to the most advanced AI models and will advise on how to use them in a government context.
Military AI
Both announcements are vague about how OpenAI’s AI models will be used by the US military. OpenAI promises in the blog that ‘the use of the technology must be in line with its own guidelines and terms of use’. This does not mean that AI can be deployed as a weapon: OpenAI limits it to ‘administrative support, improving healthcare, and enhancing defensive capabilities’.
The Department of Defense, on the other hand, writes that OpenAI, as a technology provider, will be responsible for “the development of prototypes of groundbreaking AI capabilities to address critical national security challenges in both warfare and enterprise domains”. The mention of the word “warfare” seems to somewhat point towards a broader use of AI technology.
The US military shows extensive interest in AI technology. OpenAI rival Anthropic is also opening its models to defense and intelligence services. While military collaborations used to be a sensitive topic within the technology sector, the perception around it has completely changed. Under the guise of national security, moral principles are quickly thrown overboard.
read also