On this episode of Virtual Sentiments, Kristen Collins talks with Katharine Jarmul about how transparency, open-source AI, and privacy intersect in today's debates over artificial intelligence. Together they unpack the differences between proprietary, open-weight, and truly open-source AI models. They also explore how AI companies shape user experiences through design choices that often remain invisible, why smaller task-specific models may offer important advantages over increasingly centralized frontier models, and what concepts like data sovereignty and individual autonomy mean in practice. Rather than framing AI as an inevitable force beyond public influence, Collins and Jarmul make the case that curiosity, transparency, and community participation can help people reclaim agency over the technologies shaping their lives.

Katharine Jarmul is a data scientist who focuses her work and research on privacy and security in data science, deep learning and AI. She is the author of the well-received O'Reilly book, Practical Data Privacy, and has more than 10 years’ experience in machine learning and AI, where she has helped build large-scale AI systems with privacy and security built in.

**This episode was recorded on June 24, 2026**

Show Notes:

Probably Private: A newsletter on data privacy by Katharine Jarmul

Youtube Channel | Probably Private

How AI / ML Memorization Happens: Repetition

Katharine Jarmul, Busting AI Myths and Embracing Realities in Privacy & Security

Robert McMillan, “How Heartbleed Broke the Internet — And Why It Can Happen Again” (WIRED, 2014)

Sebastian Raschka, Build a Large Language Model (From Scratch)

C. E. Shannon, “A Mathematical Theory of Communication” (The Bell System Technical Journal, 1948)

Ken Yeung, “The New Side Hustle for Graduate Students: Training AI” (The AI Economy, 2025)

Ruth Fowler, “I Work in Hollywood. Everyone Who Used to Make TV Is Now Secretly Training AI” (WIRED, 2026)

Laura Kelley, “The Work of Helping A.I. Destroy Work” (NYT, 2026)
Joel Mokyr, "The Intellectual Origins of Modern Economic Growth" (The Journal of Economic History, 2005)

If you like the show, please subscribe, leave a 5-star review, and tell others about the show! We're available on Apple Podcasts, Spotify, Amazon Music, and wherever you get your podcasts.

Follow the Hayek Program on Twitter: @HayekProgram

Learn more about Academic & Student Programs

Follow the Mercatus Center on Twitter: @mercatus