AI Chat: ChatGPT, AI News, Artificial Intelligence, OpenAI, Machine Learning Navigating AI Privacy with Jonathan Mortensen
Oct 15, 2025
Jonathan Mortensen, Founder and CEO of Confidence Security, dives into the pressing issue of AI privacy and data breaches. He shares innovative strategies like anonymization and trusted execution environments to protect sensitive data. The discussion spotlights target markets including enterprises and sovereign AI applications. Jonathan explains integration options, weighing managed services against on-premise solutions. He also highlights common pitfalls like prompt injection risks, encouraging awareness to safeguard AI security.
AI Snips
Chapters
Transcript
Episode notes
AI Raises Stakes Of Data Breaches
- AI raises stakes beyond typical breaches because trade secrets and privileged data can be exposed.
- Companies should treat AI data risks as higher-impact than standard consumer data leaks.
Recent Chat Leaks And Vendor Breach Costs
- Jonathan cites recent incidents where major providers accidentally exposed chat threads indexed by Google.
- He uses enterprise breach statistics to show third-party breaches are nearly universal and costly.
Private Compute Architecture Applied Broadly
- Confident Security implements Apple's Private Cloud Compute architecture for provable private inference.
- They combine anonymization with trusted execution environments so only the AI can decrypt and run prompts.