When AI Agents Own the Endpoint: The Security Gap No One's Watching
Sponsored by:
AI agents now execute shell commands, read files, install packages, and make network calls autonomously on endpoints - but they don't behave like humans, and they don't behave like traditional software. This session shows what makes agentic AI a fundamentally different endpoint threat: agents that break containment to complete a task, attack surfaces that shift with every weekly release, and real cases where agents bypass security controls with no malicious prompt required. If you're thinking about this through the lens of traditional endpoint security, you're already behind.
Real scenarios, real gaps, no theory.
Speakers
Yossi Pik
Co-Founder & CTO, Backslash Security
Yossi Pik is Co-Founder and CTO of Backslash Security, building security for AI-native development and agentic endpoints. He previously co-founded and served as CTO of Farmigo (acquired by GrubMarket) and was VP Engineering at SAP. He runs a team of nine AI agents for daily product, research, and content work, and writes about AI agentic security from hands-on experience building and securing agent systems.
