Wave of ChatGPT uninstallations: OpenAI reviews its agreement with the military, but acknowledges its limitations

© Mariia Shalabaieva

The agreement made with the US government also caused a stir among OpenAI employees. By aligning with their counterparts at Google, OpenAI staff published an open letter. This letter encourages startup leaders to « continue to refuse current requests from the Department of Defense to authorize the use of our models for domestic mass surveillance and autonomous elimination of people without human control ». The letter was signed by nearly 1000 people.

Also read: The creator of ChatGPT downplays AI consumption with a surprising argument

OpenAI tries to quell the controversy

Faced with growing controversy, Sam Altman, CEO of OpenAI, took a stand. The entrepreneur admits that his company should not have « rushed » in communicating about the agreement so quickly. We « were genuinely trying to defuse tensions and avoid a much worse situation, but I think it gave an impression of opportunism and negligence », explains Altman.

Here is a re-post of an internal post:

We have been working with the DoW to make some additions in our agreement to make our principles very clear.

1. We are going to amend our deal to add this language, in addition to everything else:

“• Consistent with applicable laws,…

— Sam Altman (@sama) March 3, 2026

To quell the controversy, the American businessman adds that OpenAI has insisted on « making some changes » to its agreement with the Pentagon. The agreement now includes clauses that specify that « the AI system must not be used intentionally for domestic surveillance operations targeting individuals or U.S. citizens ».

In addition, the modifications explicitly prohibit intelligence agencies, including the National Security Agency (NSA), from using ChatGPT technologies. Any « service provision to these agencies would require a new agreement », which suggests that a tailored contract for intelligence agencies is not excluded. On X, Sam Altman points out that the War Department understands the new clauses of the partnership.

« The technology is not yet mature enough for many use cases, and we don't yet fully understand all the security compromises to make. We will work on it progressively, in collaboration with the Ministry of War, relying on technical devices and other approaches », says the CEO of OpenAI.

Limited oversight

Despite the clauses added by the startup to reassure its employees and users, OpenAI would not have a say in how the AI is used by the army, reports CNBC. During an internal meeting, Sam Altman was very clear : the company has no leeway to « make operational decisions » regarding how the government uses its artificial intelligence technology.

« You may think the strike against Iran was justified and the invasion of Venezuela was a mistake. You don't have a say on that », summarizes the CEO.

Nevertheless, Sam Altman assures his employees that the Pentagon recognizes OpenAI's technical expertise and wants their input on the relevant uses of its models. The Trump administration will also allow the company to design the security system it deems appropriate, but that's where it stops. The Pentagon has clearly emphasized that operational decisions are its responsibility.

👉🏻 Follow real-time tech news : add 01net to your Google sources, and subscribe to our WhatsApp channel.