Dan Driscoll, former United States secretary of the Army, retired after conflicts with Secretary of War Pete Hegseth and has moved on to become the senior counselor for national security modernization at OpenAI. Since Hegseth became the secretary of defense, he has fired over a dozen senior Pentagon officers, leaving the Department of War with significant vacancies and appointing individuals connected to Elon Musk’s Department of Government Efficiency team to senior roles.
With the Department of War in turmoil, initiatives by Hegseth like Project Meridian and Autonomous Warfare Command have exploited unregulated artificial intelligence. Hegseth’s goals include “present the proactive vision for shaping American dominance in the future of warfare” and “expanding on the robotic systems to the Middle East in support of Operation Epic Fury; the fastest peacetime operational shift in modern military history”.
OpenAI hired Driscoll to safely modernize the Pentagon’s application of AI with the intention of “strengthening national and cyber security”. Driscoll’s main focus includes using AI to “help them keep pace by cutting through paperwork, making sense of more information and getting the right answer to the right person faster.”
OpenAI’s Day Break Program has been given $1 billion to achieve its global commitment to expand subsidized access to AI. Its Day Break Defense Network brings cyber models to businesses and frontline defenders to protect American systems and services. The Department of War has also enacted its own AI platform called GenAI.mil and will also add ChatGPT Enterprise and ChatGPT Gov to its framework to improve battle-readiness and efficiency.
The use of AI in the military is not unthinkable, as the U.S. and Israel have already been using Palantir’s Maven Smart System to identify Iranian targets. The system parses information from satellites, drones, data brokers and social media to identify people and objects. In Gaza, Israeli Defense Forces can even bypass the verification process for AI-recommended target areas when analysts operate under tight deadlines.
In July 2025, the Trump administration offered top U.S. technology companies $200 million contracts in exchange for access to their AI tools. While OpenAI accepted the government’s offer, Anthropic declined. When the Pentagon requested “full, unrestricted access to Anthropic’s models,” Dario Amodei, wary of the inaccuracies of AI, responded with reluctance to allow “mass domestic surveillance and fully autonomous weapon systems.”
After rejecting Hegseth’s ultimatum, Trump designated Anthropic as a supply chain risk to national security and prevented federal agencies from using the company’s services.
