The Pentagon Asked OpenAI for Artificial Intelligence Designed to Rarely Say No
The U.S. military’s desire for an AI tool with “minimal refusal rates” was revealed through a FOIA lawsuit filed by The Intercept. The post The Pentagon Asked OpenAI for Artificial Intelligence Designed to Rarely Say No appeared first on The Intercept .
The Pentagon solicited OpenAI to create a specialized version of its artificial intelligence that would rarely decline commands, according to documents obtained by The Intercept through a Freedom of Information Act lawsuit. This request for a "minimal refusal rates" AI tool emerged in files provided to The Intercept, which detailed the military's secretive agreements with AI companies.
OpenAI, alongside rivals Google, xAI, and Anthropic, had agreed in 2025 to develop military-focused AI prototypes for tasks like logistics, intelligence decision-making, and overall "warfighting." However, the Pentagon sought to expand these agreements and deploy them across classified U.S. computer networks, leading to a clash with Anthropic over the extent of military uses.
The contract seeking "minimal refusal rates" from OpenAI is part of a revised agreement (version "P00003"), worth up to $200 million over two years. OpenAI and the Pentagon deny agreeing to such language, asserting that the disputed document was a draft and not the final version. The Pentagon maintains that the "P00003" contract was finalized and that they never agreed to require "minimal refusal rates."
Written by urgent.news from The Intercept's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
