Practical: AI & Business News Guarding the Future: AI Privacy Insights with Jonathan Mortensen
Oct 19, 2025
In this insightful discussion, Jonathan Mortensen, Founder and CEO of Confidence Security and a Stanford PhD, delves into the complexities of AI privacy. He highlights the real risks enterprises face when handling confidential data and explains how his company pioneers provable private inference with anonymization and trusted execution environments. Jonathan discusses the various sectors that can benefit from enhanced security, the integration processes for companies, and the importance of maintaining security standards across deployment options. He also reveals common security pitfalls and how to avoid them.
AI Snips
Chapters
Transcript
Episode notes
AI Elevates Data Breach Stakes
- AI raises stakes because leaks can expose core trade secrets, not just personal data.
- Recent accidental public indexing of chats shows this risk is real and ongoing.
Publicly Indexed Chat Threads
- Major providers accidentally had chat threads indexed and made publicly searchable by Google.
- This example shows users didn't expect chats to become public and underscores practical risk.
Hardware-Based Private Inference
- Confidential computing uses anonymization plus encrypted execution to prevent data exposure.
- Trusted execution environments decrypt data only inside encrypted memory so operators can't inspect it.