Navigating the most common threats in AI-based systems

As organisations accelerate the development and deployment of AI‑based applications, the security landscape is evolving just as rapidly. AI systems introduce new attack surfaces, unique failure modes, and vulnerabilities that traditional security frameworks do not fully address. Understanding these risks – and the methods to systematically assess and mitigate them – has become essential for anyone building or integrating AI technologies.

In this webinar, we explore today’s most common threats facing AI-driven systems, from prompt-level attacks to adversarial inputs designed to manipulate model behaviour. We will also discuss how organisations can systematically evaluate AI risks using threat modelling techniques. In particular, we will explore the MAESTRO framework, which is designed specifically for identifying and analysing risks in Agentic AI‑based systems.

An AI-focused startup will present practical insights from implementing the MAESTRO framework in real-world systems, highlighting key lessons learned and challenges encountered.