Privacy Theater Is Not Privacy Engineering: What It Actually Takes to Ship Safe AI
About this episode
The episode discusses the misconceptions of AI privacy and the technical measures necessary for building safe AI systems.
Katharine Jarmul , Privacy in ML/AI Expert & Author of Practical Data Privacy , joins Hugo to unpack why most AI privacy advice is theater: and what technical privacy actually looks like when you’re shipping LLMs, agents, and multimodal systems into the real world. In this episode, we dig into how to build defensible systems in an era of AI agents and multimodal models : why system prompts (and your entire agent harness!) should be considered public by default , and why “privacy observability” is as critical as data observability for anyone building with LLMs today. Multimodal is what changes the threat model: identifiers hide in images, audio, and metadata, not just text, and the old anonymization playbook doesn’t cover it. Vanishing Gradients is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. We Discuss: * No Convenience Tax , you don’t have to trade privacy for utility: high-utility AI products can be privacy-preserving through technical controls like privacy routing and input sanitization ; * Public Prompts and Harnesses : assume any instruction or secret in a system prompt or agent harness will be…
People in this episode
Host: Hugo Bowne-Anderson
Guest: Katharine Jarmul
Topics covered
- AI privacy
- privacy engineering
- large language models
- multimodal systems
- data observability
- technical privacy
Keywords
- privacy theater
- AI agents
- system prompts
- privacy observability
- input sanitization
- data flows
- anonymization
- threat model
Mentioned in this episode
Books & works: Practical Data Privacy
More episodes of Vanishing Gradients
- Next Level AI Evals for 2026
- LLM Architecture in 2026: What You Need to Know with Sebastian Raschka
- Episode 72: Why Agents Solve the Wrong Problem (and What Data Scientists Do Instead)
- Episode 71: Durable Agents - How to Build AI Systems That Survive a Crash with Samuel Colvin
- Episode 70: 1,400 Production AI Deployments
- Episode 69: Python is Dead. Long Live Python! With the Creators of pandas & Parquet
Explore listener stats, chart rankings, contacts and more on the Vanishing Gradients podcast page.