Glenn Parham
AI, government, and national security. | Washington D.C. | glenn@glennparham.com | LinkedIn | X | Instagram
I was the Pentagon’s first generative AI technical lead, worked on AI & national security at Meta and Microsoft, and studied/taught data and computer science at UC Berkeley. Now I lead public sector at Vals AI.
I’m focused on a question I kept running into during my years in government: how do we know if AI is ready for the work we need it to do?
The only way to answer that is to test (“evaluate” or “benchmark”) these models…a lot. So we’re building some of the first benchmarks for AI child safety, economic impact, military use, geopolitical bias, cybersecurity, and more.
Our research is already cited by all of the frontier AI labs (see Zuckerberg and Sundar Pichai tweets). My job is to communicate our research to policymakers and governments here and around the world so they can make informed decisions.
Things I’m thinking about a lot these days:
How AI performs in non-English languages. “Accuracy of AI chatbots declined 16% when prompted in Spanish about voter information.” (Politico)
One in three teens has used AI companions for social interaction and relationships, including romantic interactions. (Common Sense Media)
How China has regulated AI. (CASRAI)
Recent work
Our first Vals Public Sector Summit (October 7, 2026): We brought researchers, former senior USG officials, and congressional staff together in D.C. to talk about youth wellbeing, negotiating AI with China, and what government needs from AI evaluators.
A US AI slowdown will make Chinese models more attractive (CEPA, September 21, 2026): My take on a U.S. AI slowdown, competition with China, and why we should focus on risks we can actually measure.
Media/Appearances
Negotiating AI with China: Export Controls, Competition & Cooperation (Vals AI Public Sector Summit, September 23, 2026)
Measuring AI: What Policymakers Can Learn from Benchmarks (a16z AI Policy Brief, August 2026)Bridging AI Adoption Gaps Across the Atlantic (CEPA, May 2026)
AI Evolution: The Geopolitics of the Full Stack (CEPA Forum Tech & Security Conference, October 9, 2025)
USAi Platform Breaks Down Barriers to AI for Federal Workforce (GovCIO, August 2025)
AI Diffusion and Defense Production (CSIS, April 2, 2025)
Inside the Pentagon’s Generative AI Task ForceUC Berkeley Lecture: Bringing AGI to DoD
Experience
Vals AI. Head of Public Sector (June 2026 to present). I lead our public sector work on independent AI evaluation.
CEPA. Fellow, Tech Policy Program (August 2026 to present). AI adoption, national security, and transatlantic technology policy.
Meta. Public Policy Manager, National Security (November 2025 to June 2026). I worked on AI adoption and policy for national security, getting Meta’s AI/tech into the hands of US gov and our allies.
Chief Digital & AI Office / Department of Defense. July 2022 to February 2025. The first Generative AI Technical Lead for DoD, working on Task Force Lima.
Defense Digital Service / Department of Defense. Head of Engineering (July 2022 to November 2024). I led engineering for the Pentagon’s SWAT team of nerds, building AI and software for defense missions, including tools for adapting language models to DoD use cases.
Pundit Analytics. Founder (August 2018 to August 2020). Acquired by Bullpen Strategy Group.Microsoft. Software Engineering Intern (May to August 2018). I worked on Windows IoT in Azure, deploying image recognition models that ran directly on devices.
Political Computer Science @ Berkeley. Founder (January 2017 to August 2018). I started a Berkeley student organization focused on using technology to tackle political and policy problems.- Undergraduate @
UC Berkeley: Studied data & computer science. Was also a Data Science Seminar Instructor. (August 2016 - December 2019)
Awards and honors
Rising Star Award (GovCIO, February 2025)
Certificate of Appreciation (DoD Chief Digital & AI Officer, December 2024)
Flag Flown Over the Pentagon (Defense Digital Service, November 2024)