Companies are at increased risk of violating Europe’s tough privacy laws because boards have little awareness of the different types of AI or their inherent risks. While many organizations are more familiar with predictive and conversational AI tools, such as chatbots that carry little or no discernible risk, they have not grasped that generative AI—and especially agentic AI—are different animals.
The former has the capability to produce new text and content based on historical data; the latter can reason, plan, self-evaluate, and execute actions on its own without human input or oversight. In short, where generative AI generates, agentic AI acts. Yet despite the vast differences in capability and autonomy, experts warn that many companies continue to treat agentic systems as if they were standard chatbots, overlooking the need for active human oversight and clear constraints on the technology’s ability to act within the company’s systems.
