Gladstone AI

Gladstone AI was founded in 2022 by Jeremie Harris and Edouard Harris, brothers with backgrounds in physics and AI entrepreneurship who previously co-founded an AI training company that went through Y Combinator and was later acquired. Inspired by the release of GPT-3 in 2020 and their direct relationships with researchers at frontier AI labs including OpenAI, they recognized the potential for catastrophic risks from advanced AI and decided to focus on bridging the gap between technical AI development and government preparedness.

The company operates three main service lines. The Foundations of AI course provides non-technical executive education on AI capabilities, risks, and national security implications, serving institutions including the U.S. Air Force, Space Force, multiple combatant commands, Lockheed Martin, and Arizona State University. The AI Observatory is a software platform for rapid prototyping, testing, and evaluation of large language model applications within secure government environments, deployed on Microsoft Azure Government Cloud at IL5 classification, and used by teams within the Department of the Air Force. The AI Policy Support service works directly with U.S. government agencies to craft national security safeguards for advanced AI.

Gladstone's most prominent achievement is the report "Defense in Depth: An Action Plan to Increase the Safety and Security of Advanced AI," commissioned by the U.S. State Department's Bureau of International Security and Nonproliferation in October 2022 and completed in February 2024. Developed over 13 months with input from more than 200 stakeholders including frontier AI lab researchers, cybersecurity experts, and senior national security officials across the U.S., U.K., and Canada, the 280-page report was the first U.S. government-commissioned assessment of catastrophic AI risks on the path to AGI. It warned that advanced AI could pose an extinction-level threat to humanity and recommended a "Defense in Depth" strategy encompassing interim safeguards, early warning systems, new regulatory agencies, and international coordination.

In April 2025, Gladstone published a second major national security report, "America's Superintelligence Project," based on over 100 interviews with special forces operators, intelligence professionals, and AI researchers, and physical security tours of major hyperscaler data centers. The report identified critical vulnerabilities in AI lab security and data center infrastructure, warning of CCP penetration of frontier labs and recommending a government-backed national superintelligence project.

Gladstone is an inaugural member of the Department of Commerce's AI Safety Consortium (AISIC) and has trained hundreds of Department of Defense personnel including senior executives and flag officers. The company deliberately operates without venture capital or institutional investors, funding itself entirely through government contracts, including a seven-figure Air Force contract to build and scale the AI Observatory.