OpenAI amends United States military contract following public backlash and surge in app uninstalls

Sam Altman adds guardrails to OpenAI's Pentagon contract, banning domestic surveillance after uninstalls of ChatGPT jumped 295% following military deal.

By: AXL Media

Published: Mar 3, 2026, 8:18 AM EST

Source: The information in this article was sourced from Chris Vallance and Laura Cress

Amending the Pentagon agreement

OpenAI is moving to add specific legal language to its contract with the United States government after the initial deal was labeled "opportunistic and sloppy" by its own Chief Executive. Sam Altman stated on Tuesday, March 3, 2026, that the new amendments will explicitly prohibit the use of OpenAI systems for the domestic surveillance of U.S. persons. Furthermore, intelligence agencies like the National Security Agency (NSA) will be barred from utilizing the company's technology without further contract modifications. Altman admitted that the company rushed the announcement on Friday in an attempt to de-escalate concerns following the fallout between the Department of Defense and rival AI firm Anthropic.

Market reaction and user backlash

The decision to work on classified military operations has triggered a massive reaction from the public. Day-over-day uninstalls of the ChatGPT mobile app spiked to 295 percent on Saturday, a sharp departure from the typical 9 percent churn rate. Simultaneously, Anthropic’s Claude AI has ascended to the top ranking in Apple’s App Store. This shift is particularly notable as the Trump administration recently blacklisted Anthropic after the firm refused to allow its models to be used in the development of fully autonomous "killer" robots. The move suggests a growing consumer preference for "safety-conscious" AI actors, even as they are marginalized by official government policy.

The role of AI in the 2026 conflict

Artificial intelligence has become a core component of modern military operations, particularly in the ongoing US-Israel war with Iran. Companies like Palantir provide data analytics tools such as "Maven" to NATO and the U.S. military, which integrate satellite data and intelligence reports. These systems utilize commercial large language models to process vast amounts of information and facilitate more efficient military decisions. While OpenAI aims to fill the gap left by Anthropic, experts like Professor Mariarosaria Taddeo of Oxford University warn that removing the most safety-focused companies from the Pentagon's "room" could lead to higher risks in autonomous weapon development.