Multi-Modal Agentic AI Cybersecurity Threats
đïž 3 June, 2025
đ 17:30 â 19:30
đ The Park Forskaren
The future is undoubtedly autonomous, and the era of Agentic AI is here.
As frontier model capabilities are evolving, however, they are doing so much faster than our safeguards, leaving powerful AI systems exposed to major hacking risks.
In this talk, weâll dive deep into the bleeding edge of security vulnerabilities in multi-modal Agentic AI systems. Youâll understand threat models and get a front-row look at how adversaries can target an agentâs reasoning brain, exploit its memory, and manipulate the tools it uses to interact with the world, across language, vision, audio, and code.
About the Speaker
Ali Leylani is the Chief AI Officer of Echo Alpha and a seasoned AI security professional.