Ioannis Agrafiotis is a Senior Research Associate in the Department of Computer Science at the University of Oxford, with more than 17 years of experience conducting research and consulting in cybersecurity. He has also worked for 6 years at the European Union Agency for Cybersecurity (ENISA) as an expert in capacity building. Ioannis has extensive experience advising governments and organisations on cybersecurity-related topics (governance, technical controls, and the use of AI) and has delivered training to IT and cybersecurity professionals.
As organisations rapidly deploy LLMs and agentic AI systems, attackers are discovering new ways to exploit them. In this session, Ioannis Agrafiotis, Senior Research Associate at the University of Oxford and former cybersecurity expert at the European Union Agency for Cybersecurity (ENISA), will explore the most significant security threats facing AI today, drawing on real-world incidents including the EchoLeak attack targeting Microsoft Copilot and the compromise of McKinsey's Lilli AI assistant. Through audience discussion and practical case studies, the session will examine how these attacks work, the risks they expose, and the controls needed to build secure and resilient AI systems.
• Understanding emerging attacks against LLMs and AI agents.
• Addressing critical AI governance and security risks.
• Applying practical controls for secure AI deployment.
Check out the incredible speaker line-up to see who will be joining Ioannis.
Download The Latest Agenda