While not technically new, agentic AI is currently one of the most buzzed-about developments in artificial intelligence, with Gartner recently naming it on of 2025’s top technology trends. Unlike generative AI, which merely creates output in response to a prompt, agentic AI can act on the user’s behalf to fulfill multi-step requests. Agentic AI can autonomously execute tasks, employ “thinking” decision making and engage third-party applications to facilitate task completion.
Agentic tools face many of the same security challenges as their generative counterparts, but their ability to act autonomously presents an additional layer of governance challenges and unique security risks, primarily regarding ethics and scope of utilization.
Reprinted with permission from Risk Management Magazine. Copyright 2025 Risk & Insurance Management Society, Inc. All rights reserved.
Click here to view the full article