Sysdig launches AI Defense to check and block risky AI agent actions in real time
On November 4, security company Sysdig made Sysdig AI Defense generally available. It checks every action an AI agent takes, from the laptop to the cloud, and uses company policy to allow it, ask the user to confirm, or block it; it supports macOS and Windows for now. My take: once AI moves from answering questions to actually doing things, security can't just be an after-the-fact audit. Before an agent touches company systems, decide what it may and may not do, and give it as few permissions as you can.