The big picture: OpenAI has backtracked on its controversial agreement with the Pentagon following significant backlash from users and privacy advocates. On Monday, CEO Sam Altman acknowledged that the deal "looked opportunistic and sloppy" and promised additional safeguards to prevent government use of the technology for surveillance of US citizens.
On X, Altman acknowledged that OpenAI should have taken more time to address the "super complex" issues surrounding privacy and data security before rushing to finalize the agreement with the Pentagon. He added that the company learned a valuable lesson from the controversy, one that will help guide better decision-making when handling higher-stakes partnerships in the future.
Here is re-post of an internal post:
– Sam Altman (@sama) March 3, 2026
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,…
Detailing the revisions, Sam Altman said the agreement has been amended to include clauses prohibiting the government from using the company's AI software to "deliberately" track, surveil, or monitor US citizens. He emphasized that the updated deal complies with all applicable federal laws, including the Fourth Amendment, the National Security Act of 1947, and the Foreign Intelligence Surveillance Act of 1978.
Despite reworking the contract to meet legal requirements, Altman declined to apologize, maintaining that the original agreement did not violate the Constitution. He said he would rather go to jail than follow an unconstitutional order, and insisted he would never have signed the deal if it had been unlawful.
Social media users, however, are skeptical. Critics argued that the inclusion of the word "deliberate" leaves loopholes that could allow the technology to be misused. Political researcher Tyson Brody noted that the language may permit the government to collect private data on US citizens under the guise of "incidental collection," potentially rendering such practices legally permissible.
Extremely careful and concerning language. Hard not to read as admitting to an AI dragnet.
– tyson brody (@tysonbrody) March 3, 2026
"intentionally" and "deliberate" - so Americans *will* be swept up in this data, but the government can claim "incidental collection" and thus legal. See the NSA and section 702. https://t.co/Tseita4D8m
OpenAI announced its agreement with the Department of Defense last week after Donald Trump terminated the contract with Anthropic following the company's refusal to comply with Defense Secretary Pete Hegseth's request to remove safeguards restricting the use of its AI technology for mass domestic surveillance and autonomous weapons.
Critics say OpenAI secured the contract after agreeing to Hegseth's demands, which were described by opponents as overly strict and ethically concerning. Following the amendment of the agreement to incorporate the so-called safeguards, activists argued that the revised language still permits "unintentional" mass surveillance and large-scale data collection. They also contend that the policy does not adequately address risks associated with AI-powered autonomous weapons systems.