OpenAI CEO Sam Altman admitted he was hasty when announcing the cooperation agreement between OpenAI and the US Department of Defense, in the context that the company faced a wave of strong opposition from users and observers.
In a post on X social network on March 3, Mr. Altman said OpenAI is revising the agreement with the US Department of Defense (DoD), adding clauses specifying the limits on the use of AI technology.
The most notable point is the clear ban on using OpenAI's AI system to monitor citizens and US nationals within the country.
According to the supplement, AI models must comply with the Fourth Amendment of the US Constitution, the National Security Act 1947 and the FISA Act 1978, and must not be intentionally used to track, monitor or collect personal information from commercial data sources for domestic surveillance purposes.
Mr. Altman also emphasized that OpenAI models will not be deployed to intelligence agencies such as the National Security Agency (NSA) under the current framework. Any service for these agencies must have additional contract revisions.
The above moves took place after public opinion erupted in controversy about whether AI could be used for military purposes, including mass surveillance or support for autonomous weapons.
Previously, OpenAI's rival Anthropic was also entangled in tension with the Pentagon when CEO Dario Amodei rejected some data-related requests.
According to US media, the Department of Defense once proposed the possibility of analyzing large volumes of commercial data such as geolocation and web browsing history.
Tensions escalated when Defense Secretary Pete Hegseth announced the termination of cooperation with Anthropic, even labeling it "supply chain risk to national security", a term rarely applied to US businesses.
In that context, OpenAI's quick agreement with DoD has caused a part of users to react strongly, with calls to "Abolishing ChatGPT" appearing on forums.
Mr. Altman admitted: "We should not rush to announce it. The issue is very complex and needs to be conveyed more clearly.
However, OpenAI CEO affirmed that he is determined to protect civil liberties and emphasized that protecting the interests of the American people is a top priority.
He also reaffirmed the long-standing view that AI should not be used to make combat decisions.
According to Mr. Altman, current systems are not reliable enough for such situations, although they can support analyzing large volumes of intelligence reports.
OpenAI plans to hold a full-time employee meeting to answer questions about the agreement with the Ministry of Defense. Mr. Altman also said that the company will continue to add technical and legal protection measures to balance between cooperation with the government and protection of personal rights.
