Guest: Peter Holcomb, IT Security Leader, Fractional CISO, and Founder of Optimo IT
Host: Ryan Estes
Presented by: Kitcaster.com
Key Takeaways:
- The Productivity vs. Privacy Dilemma — AI can drive massive productivity gains—but companies must weigh those gains against serious privacy, security, and ethical risks.
- The Real Risks of AI Right Now — Algorithmic bias, data breaches, hallucinations, and poor access controls are among the top concerns. Always sanitize inputs and outputs.
- Stop Feeding Sensitive Info to Chatbots — Don’t input PII or PHI into commercial LLMs unless using secure, enclaved systems.
- RAG + LLMs = Security Headaches — RAG improves outputs but opens up new API and access control vulnerabilities.
- Time Machine Hypothesis — Data input today could resurface in dangerous ways tomorrow without stronger regulations.
- Guardrails Are the New Gold — Autonomous agents are rising. Governance and real-time red teaming are critical.
- Bots vs. Bots — Cybersecurity is evolving into automated battles between defensive and offensive AI agents.
- Tools to Know —
- Optimo IT’s Governance Approach — Risk assessments, red teaming, and guardrails to secure proprietary AI use.
- The Human Element — AI will always need humans in the loop to keep things aligned, ethical, and on track.
Mentioned Links:
Next Week:
We’ll bring Peter back for Part 2, assuming the AI overlords let us finish the conversation this time. Until then—stay curious, stay cautious, and don’t feed your social security number to Skynet.
__
ryanestes.info
The pod on TikTok
The pod on YouTube
The pod on Spotify
The pod on Apple
Subscribe to the newsletter here, for AI news, tools, and exclusive promos for founders.
