OpenAI Deepens Pentagon Partnership
OpenAI is giving the Pentagon access to its latest AI models for classified military use, a move that follows the government's ban on rival Anthropic. CEO Sam Altman defended the deal, citing "strict guardrails," but critics argue OpenAI caved to military demands on surveillance, blurring the line between civilian and military AI.
The deal materialized just hours after President Donald Trump ordered a government-wide ban on technology from rival AI company Anthropic. The ban followed Anthropic's refusal to remove safeguards that would prevent its AI, Claude, from being used for domestic mass surveillance or in fully autonomous weapon systems without human oversight. In response to Anthropic's stance, Defense Secretary Pete Hegseth designated the company a "supply chain risk to national security," a label historically applied to foreign adversaries like Huawei. This designation effectively bars any contractor or supplier doing business with the U.S. military from also working with Anthropic. OpenAI states its agreement with the Pentagon includes similar "red lines" to those Anthropic advocated for: no use for mass domestic surveillance and no deployment in autonomous weapons systems that operate without human control. The company says its safeguards are ensured through a cloud-only deployment, which prevents the AI models from being run on edge devices that could be used for autonomous weapons, and by keeping cleared OpenAI personnel involved in the process. This partnership is part of a broader, long-standing, and sometimes contentious relationship between Silicon Valley and the U.S. military. A notable precedent is Project Maven, a Department of Defense initiative started in 2017 to use AI for analyzing drone footage. Google was initially a major partner but withdrew in 2018 after significant employee protests against the company's involvement in military technology. The Pentagon's push for "frontier models" from companies like OpenAI and Anthropic highlights a strategic priority to integrate the most advanced AI into military operations for tasks like data analysis and target identification. These powerful models are seen as a necessity for maintaining a technological advantage. Despite OpenAI's assurances, some critics argue the contractual language may contain loopholes, particularly concerning the collection of commercially available bulk data on Americans. The debate centers on whether existing law offers sufficient protection against potential AI-powered surveillance. The agreement has sparked considerable backlash, leading to a "Cancel ChatGPT" movement on social media and a surge in downloads for Anthropic's Claude, which hit number one on Apple's App Store following the controversy. OpenAI CEO Sam Altman acknowledged the deal was "definitely rushed" and that the "optics don't look good."
