📄 中文摘要
OpenAI 发布了 Operator，这是一种能够使用计算机、浏览网站、点击按钮和自动化任务的 AI 代理。与此同时，Anthropic 也推出了 Claude 计算机使用功能，Claude 能够与屏幕互动、执行操作并自动化工作流程。这两种技术均已具备生产就绪的能力，能够自动化关键业务流程。然而，它们的发布缺乏合规框架，导致团队在使用这些代理进行工作时，合规官员提出了无法回答的问题。
OpenAI 操作员与 Claude 计算机使用：缺乏审计追踪的生产就绪代理
📄 English Summary
OpenAI Operator and Claude Computer Use: Production-Ready Agents Missing Audit Trails
OpenAI has released Operator, an AI agent capable of using your computer, navigating websites, clicking buttons, and automating tasks. Similarly, Anthropic has introduced Claude Computer Use, which allows Claude to interact with your screen, perform actions, and automate workflows. Both technologies are production-ready and can automate critical business processes. However, they launched without a compliance framework, raising questions from compliance officers that teams cannot answer while automating work with these agents.