Securing AI Applications: Understanding the New Attack Surface

AI hasn’t just changed how we build, it has fundamentally reshaped the attack surface. Today’s AI applications are dynamic, interconnected systems spanning models, data pipelines, agents, and cloud infrastructure, introducing new and often invisible paths for risk.

In this session, Snegha Ramnarayanan at Wiz, examines where risk actually exists across modern AI applications and why traditional security approaches fall short. Drawing on how leading teams are evolving their thinking, the session explores what it takes to gain end‑to‑end visibility, prioritise real risk, and protect AI systems across code, cloud, and runtime.

Key takeaways:

• Where risk sits across modern, interconnected AI applications – and why attack paths are often hard to see
• Why traditional security approaches break down in AI‑driven environments
• How leading teams approach visibility, risk prioritization, and protection across the AI stack