Today, almost all of our initial security alerts are triaged by intelligence before humans are looped in. This helps reduce toil for defenders, improves response time, and lets humans spend time where their skills are most leveraged—in discernment, judgement, and applied expertise.
OpenAI已经把AI部署为安全运营的第一道防线：机器处理初级分类，人类只介入需要判断力的高价值决策。这正是EP.99「分级授权成为默认」的核心逻辑——AI拿走执行层，人类保留决策层，权责边界通过技术设计而非人工约定来划定。