Risky Bulletin Sponsored: What AI workloads mean for Cloud security
Jan 11, 2026
Toni de la Fuente, Founder and CEO of Prowler, shares insights on the evolving landscape of cloud security as it intersects with AI workloads. He discusses unique security challenges posed by AI, highlighting how traditional tools may fall short. Toni delves into attack paths in cloud environments, stressing the importance of scanning models and applying OWASP guidance to mitigate risks. He also emphasizes the architectural complexities introduced by AI, drawing parallels to the early web era, and recommends using open-source tools to tackle emerging threats.
AI Snips
Chapters
Transcript
Episode notes
AI Is Part Of The Cloud OS
- AI workloads are tightly integrated with cloud infrastructure because data, GPUs, and APIs mostly live in the cloud.
- Securing AI therefore requires securing the cloud platform and new AI-specific services and components.
Apply Least Privilege And Guardrails
- Configure AI services and surrounding resources with least privilege and proper guardrails to avoid exposing sensitive data.
- Protect API endpoints, storage, and keys and enable controls that prevent unauthorized access and data leaks.
AI Adds New Architectural Complexity
- AI adds components and complexity beyond familiar cloud services, changing authentication and data access patterns.
- You must decide whether models talk to data directly or through limited APIs to reduce risk.