AI model providers are increasingly stepping back from enforcing guardrails, putting the responsibility for AI security squarely on enterprises. But most organizations don't yet have the visibility to meet that responsibility, facing a blind spot across the broader ecosystem of AI systems already operating in their environments. Closing that gap requires unified visibility across both AI systems and the cryptographic infrastructure they touch, so security teams can assess risk and act on it in one place.
Segment Resources:
https://www.sandboxaq.com/press/sandboxaq-report-ai-adoption-outpacing-ai-security https://www.sandboxaq.com/press/sandboxaq-launches-aqtive-guard-ai-spm-to-stop-the-rapid-spread-of-shadow-ai-across-the-enterprise
Visit https://securityweekly.com/sandboxaqrsac to discover how enterprises are taking control of their AI security with AQtive Guard AI-SPM by SandboxAQ.
Read the interview summary from SC Media here: SandboxAQ’s Marc Manzano: The onus for AI security is on the enterprise
- 0:00 - Introduction to RSAC 2026 & AI Guardrails Discussion
- 0:24 - Do AI Guardrails Even Exist Yet?
- 0:43 - New Security Challenges Introduced by AI
- 01:28 - Speed & Asymmetry Between Attackers and Defenders
- 02:14 - Why Attackers Move Faster Than Security Teams
- 02:23 - The Rise of Shadow AI in Enterprises
- 03:38 - AI Supply Chain Risks & Hidden Vendor AI Usage
- 03:40 - Data Security Challenges with AI Prompts & Semantics
- 03:54 - Malicious AI Models & Open Source Threats
- 04:43 - Lack of Traceability Across AI Agents
- 05:30 - AI as a New Insider Threat Risk
- 06:06 - AI-Powered Exploits & Zero-Day Speed Advantage
- 06:55 - Why Traditional Patch Windows No Longer Work
- 07:52 - Keeping Security Systems at Machine Speed
- 08:23 - How to Detect Shadow AI in Your Environment
- 09:38 - Moving from Inventory to Risk-Based AI Visibility
- 09:54 - Expanding AI Risks Across the Supply Chain
- 11:21 - Quantum Computing & AI Security Intersection
- 11:52 - Post-Quantum Cryptography vs Quantum Cryptography
- 12:58 - How CISOs Should Handle AI + Quantum Threats
- 13:16 - Building Separate AI and Quantum Security Programs
- 14:03 - Automating Identity & Access for AI Agents
- 15:44 - Final Thoughts on AI Security Strategy
Dr. Marc Manzano leads the cybersecurity group at Sandbox. His current research interests include post-quantum cryptography, lightweight cryptography, fully-homomorphic encryption, the intersection between machine learning and cryptanalysis, performance optimizations of cryptographic implementations on a wide range of architectures, and quantum algorithms. Manzano holds a Ph.D. in Computers Network Security, which he earned under the supervision of the University of Girona (Spain) and Kansas State University (United States). He earned an MSc in Computer Science from the University of Girona (Spain), while he did research stays at UC3M (Spain) and at DTU (Denmark). He initiated his research career while finalizing his BSc in Computer Engineering at Strathclyde University (UK).
Over the past ten years, Manzano has led the development of many secure cryptographic libraries and protocols. Manzano was formerly a Senior Staff Software Engineer at Google, and before that, he was the Vice President of the Cryptography Research Centre at the Technology Innovation Institute, a UAE-based scientific research center. Prior to that, he held several positions where he was responsible for implementing pivotal cryptographic components of a variety of secure communication products, including an electronic voting platform.
