Agentic AI: Cyber risk, data security and AI agent threats

Episode 3: AI agents expand what organisations can do, and what attackers can target. AI-enabled phishing, deepfakes, prompt injection, model theft, expanded attack surfaces, nation-state agentic attacks: the threat landscape is evolving faster than most organisations' defences.   How do you build AI agent infrastructure that is genuinely secure? What does regulatory compliance look like in this context? And how do you turn risk management into a trust-building advantage?  Alex Shandro and Anna Rudawski close this three-part series by examining how AI agents elevate cybersecurity and privacy risks — and what organisations need to do about it. They cover governance frameworks, practical controls and the regulatory expectations businesses must meet, offering a clear-eyed view of how to accelerate AI adoption without compromising security.  Listeners will leave with a concrete understanding of the threat landscape facing AI agent deployments — and the governance and technical controls needed to manage it.  Agentic AI explores the legal, regulatory and commercial implications of AI agents in business. In each episode, our specialists decode the risks, regulatory frameworks and practical steps organisations need to deploy AI agents with confidence. The series is designed for legal, compliance, risk and technology leaders navigating the fast-moving frontier of autonomous AI.