As AI systems evolve from simple chatbots into autonomous agents capable of making decisions and executing tasks, the security risks grow dramatically. In this episode, the discussion explores how frameworks from the OWASP Generative AI Security Project are helping organisations understand and defend against emerging threats in agent-based AI systems. The conversation covers key risks such as prompt injection, agent privilege abuse, and insecure tool usage, while highlighting the OWASP Top 10 for Agentic Applications as a practical guide for securing modern AI deployments. Listeners gain insight into why traditional security models struggle with autonomous AI, and how developers, security teams, and organisations can begin building safer AI systems from the ground up.
