Title:

Securing the AI Lifecycle: Trust-by-Design Approaches to Adversarial Threats

Keynote speaker:

Prof. Christos Xenakis, University of Piraeus, Piraeus, Greece

Prof. Christos Xenakis is Professor in the Department of Digital Systems at the University of Piraeus, Greece, and Director of the Postgraduate Programme “Cybersecurity and AI Technologies”. He holds a B.Sc. in Computer Science (1993), an M.Sc. in Telecommunications and Computer Networks (1996), and a Ph.D. (2004) from the Department of Informatics and Telecommunications, University of Athens. His experience includes telecom systems development and long-standing involvement in the Communication Networks Laboratory at the University of Athens. He has coordinated multiple EU-funded cybersecurity projects, including AIAS, ANTIDOTE, NITRO, SECONDO, INCOGNITO, ReCRED, CUREX, and SealedGRID, and has served as technical manager of cPAID and UINFC2. He is a member of the steering committee of the European Cyber Security Challenge (ECSC) and leads the Hellenic Cyber Security Team. He serves on the editorial boards of Computers & Security, Computer Communications, and The Computer Journal. His research focuses on systems, network, and application security, with more than 130 peer-reviewed publications.

Abstract:

As Artificial Intelligence (AI) technologies grow exponentially, they are transforming domains from smart cities to medical devices. At the same time, this rapid evolution significantly expands the attack surface, exposing AI systems to adversarial threats such as poisoning, evasion, inference, and extraction attacks.

This talk will present an AI defence framework designed to strengthen AI systems against this wide range of attacks. The framework integrates multiple defensive mechanisms, including Generative Adversarial AI, AI-assisted Intrusion Detection and Prevention Systems, Risk Management for AI, Data Fabric, Meta-SIEM, and an Adversarial AI Cyber Range.

It will also explain how the platform applies the MLPrivSecOps methodology, embedding security-, privacy-, and trust-by-design principles throughout the AI lifecycle. The talk will walk through the architecture, methodologies, and core components of the framework, highlighting its scalability, robustness, and alignment with emerging ethical AI principles.