📄 中文摘要
在过去十年中，人工智能治理通常处于被监管系统之外，政策制定、审查、模型批准和审计等活动在事后进行。这种分离在 AI 作为工具时有效，能够按需生成预测或建议。然而，随着 AI 的发展，这一假设正在崩溃。治理需要向系统内部转移，以应对 AI 的自主性和复杂性，确保其行为符合伦理和法律标准。内嵌治理机制将有助于实时监控和调整 AI 系统的决策过程，从而提高透明度和责任感。
自主 AI 的控制平面：治理为何必须内嵌于系统中
📄 English Summary
Control Planes for Autonomous AI: Why Governance Has to Move Inside the System
For most of the past decade, AI governance has existed outside the systems it was intended to regulate, with policies, reviews, model approvals, and audits conducted after the fact. This separation worked while AI functioned merely as a tool, generating predictions or recommendations on demand. However, this assumption is breaking down as AI evolves. Governance must shift inside the system to address the autonomy and complexity of AI, ensuring its actions align with ethical and legal standards. Embedding governance mechanisms will facilitate real-time monitoring and adjustment of AI decision-making processes, enhancing transparency and accountability.