The Internet Is About to Get WAY More Dangerous (Here’s Why)

The video discusses the evolving capabilities of automation, specifically focusing on agentic AI and how agents are moving beyond simple API interactions to perform full computer operations, mimicking human screen interaction. This development impacts areas such as evidence collection, access reviews, and routine security tasks, while simultaneously introducing new inherent security risks.

Topics covered include implementing safeguards for new automation workflows, emphasizing read-only actions and sandboxed environments for initial gains. It prompts consideration of how tools can detect anomalous access requests and maintain human oversight through rapid approval steps. Since these agents can browse systems, the discussion touches upon preventing unauthorized data exfiltration.

Jeremy Epling, Chief Product Officer at Vanta, provides insight on leading product strategy for their trust management platform, focusing on helping organizations automate security and compliance.