Big Technology Podcast Is Generative AI a Cybersecurity Disaster Waiting to Happen? — With Yinon Costica
103 snips
Sep 24, 2025 Yinon Costica, co-founder and VP of Product at Wiz—a cloud security company acquired by Google—dives into the cybersecurity implications of generative AI. He reveals how AI creates new software vulnerabilities and discusses alarming trends like ‘vibe coding.’ Yinon highlights the asymmetry faced by defenders against automated attacks and the urgent need for robust security measures. He also touches on the risks of hacking autonomous vehicles and the importance of democratizing security across organizations to fend off emerging threats.
AI Snips
Chapters
Transcript
Episode notes
AI Is A New Software Stack
- AI represents a whole new software stack with prompts, models, and infrastructure interacting.
- Like any software, AI components have bugs and can expose critical vulnerabilities such as RCEs.
Infrastructure Is The Primary Attack Surface
- AI apps sit on existing cloud infrastructure that still suffers misconfigurations and exposed storage.
- Threat actors will target those misconfigurations to exfiltrate training and sensitive data.
Embed Security Rules Into Code Generation
- Instruct code generators with security rules and least-privilege practices when producing code.
- Maintain ownership and readiness to patch and operate AI-generated code over time.