96% of IT pros say AI agents are a security risk, but they’re deploying them anyway

The same capabilities that make agents much more powerful than traditional chatbots also make them much bigger potential liabilities.