Report: AI Power Struggle Erupts at Pentagon
A recent report highlights internal power struggles at the Pentagon regarding its AI partnerships with major tech firms. The disputes reportedly intensified following a deadly raid, causing top AI companies to weigh the ethical and operational implications of their continued collaboration with the U.S. military. The situation underscores the growing tension at the intersection of AI, ethics, and defense.
- The recent disputes reportedly center on a U.S. military operation in Venezuela that used Anthropic's Claude AI model for real-time data processing during a raid to capture Nicolás Maduro. The incident has jeopardized Anthropic's relationship with the Pentagon, as the company's terms of use prohibit its technology from being deployed for violent purposes or weapons development. - In July 2025, the Pentagon's Chief Digital and AI Office (CDAO) awarded contracts worth up to $200 million each to Google, xAI, Anthropic, and OpenAI. The objective is to leverage their advanced models to develop "agentic AI workflows" to support various national security missions. - A central point of contention involves the Pentagon's push for these companies to deploy their AI tools on classified networks without the standard safeguards that are in place for public users. While Anthropic's model is reportedly the only one currently available in some classified settings, officials have bristled at its usage restrictions. - This isn't the first time such conflicts have arisen. In 2018, thousands of Google employees signed a petition protesting the company's involvement in Project Maven, a program using AI to analyze drone surveillance footage. The backlash led Google to not renew the contract and to release a set of ethical principles that included a ban on developing AI for use in weapons. - After Google backed out, data analytics and software firm Palantir Technologies took over the Project Maven contract. In May 2024, Palantir received a $480 million contract from the U.S. Army to significantly expand the use of the "Maven Smart System" for data analysis and target detection. - The broader strategic context is the Pentagon's "AI Acceleration Strategy," officially released in January 2026. This policy directs the Department of Defense to become an "AI-first" warfighting force and to remove bureaucratic barriers to rapidly experiment with and deploy leading commercial AI models. - The divide in philosophy is personified by key leaders. Anthropic CEO Dario Amodei has publicly expressed caution regarding the use of AI in autonomous lethal operations. In contrast, Secretary of War Pete Hegseth stated in January that the department would not "employ AI models that won't allow you to fight wars".
