Analysis Argues AI-Driven SOCs Can Autonomously Resolve 98% of Security Alerts, Leaving Under 2% for Humans

A new analysis contends that AI-powered security operations centers (SOCs) can investigate every incoming alert and resolve roughly 98% of them autonomously, reserving fewer than 2% for human review—provided defenders deploy the technology for the right job. The argument, laid out in a piece examining the future of Managed Detection and Response, reframes the debate over tools like Claude and Codex: the models are not the bottleneck, but rather how teams draw the line between machine and human judgment. Rethinking MDR as Attackers and Defenders Embrace AI makes the case that traditional MDR, built around human analysts triaging a prioritized subset of alerts, is failing as attackers themselves adopt AI.