Afternoon Cyber Tea with Ann Johnson Security and Agentic AI
5 snips
Jun 3, 2025 Yonatan Zunger, Corporate Vice President and Deputy Chief Information Security Officer for Artificial Intelligence at Microsoft, shares his expertise in AI security. He discusses how generative AI alters the cybersecurity landscape, highlighting both opportunities and risks. Zunger emphasizes treating AI as a valuable junior employee under human oversight. He promotes the importance of robust security frameworks and proactive governance in AI development, and expresses optimism about AI's potential to enhance persistent cybersecurity defenses.
AI Snips
Chapters
Transcript
Episode notes
AI as a Fallible Junior Employee
- People often misunderstand AI as infallible when it is prone to mistakes like a junior employee.
- Effective use of AI requires human expertise and oversight to correct its errors.
Expert Corrects AI Mistakes
- An expert corrected a colleague relying solely on Copilot, showing AI can confidently give incorrect answers.
- This highlights AI's limitations and the need for expert validation in decision-making processes.
Building with AI Like a Junior
- Treat AI like a clever but fallible junior employee to build robust business processes.
- Combine AI with human checks to account for AI's frequent mistakes and improve reliability.