China recently released its AI Safety Governance Framework 3.0, a must-read for AI governance professionals, especially given the recent increase in AI safety incidents.中国近期发布了《人工智能安全治理框架 3.0》。鉴于近期人工智能安全事件频发，该框架对于人工智能治理专业人士而言是必读之作。
Published in both Chinese and English, the framework addresses new trends in AI development and responds to new challenges in AI safety governance.该框架同时发布了中英文版本，旨在应对人工智能发展的新趋势，并回应人工智能安全治理领域出现的新挑战。
Version 3.0 of the Framework follows the core approach of risk classification > technical response > comprehensive governance and is based on the following principles:《框架》3.0 版本遵循“风险分级 > 技术应对 > 综合治理”的核心路径，并基于以下原则：
Be inclusive and prudent to ensure safety包容审慎，确保安全
Risk-oriented and agile governance风险导向，敏捷治理
Integrating technology and management for coordinated response技术与管理结合，协同应对
Promoting openness and cooperation for joint governance and shared benefits开放合作，共治共享
Ensuring trustworthy application and preventing loss of control确保应用可信，防止失控
Below are some of the recommended governance measures described in Framework 3.0 to address AI safety risks:以下是《框架》3.0 中针对应对人工智能安全风险所提出的部分治理建议措施：
Improving top-level design of AI safety governance完善人工智能安全治理顶层设计
Enhancing AI safety governance capacity提升人工智能安全治理能力
Establishing a flexible, dynamic, and controllable AI regulatory sandbox mechanism建立灵活、动态、可控的人工智能监管沙盒机制
Strengthening measures for AI safety management强化人工智能安全管理措施
Deepening consensus on AI safety governance深化人工智能安全治理共识
Most companies are unprepared for the rising number of AI safety incidents, which are expected to get worse in the coming months and years.大多数公司对日益增多的人工智能安全事件准备不足，而这些事件预计在未来几个月和几年内会愈演愈烈。
This is the time to review, update, and strengthen internal AI governance policies, risk assessments, and response mechanisms, especially those focused on safety.现在正是审查、更新和加强内部人工智能治理政策、风险评估及应对机制（尤其是侧重于安全方面的机制）的时候。
This is one of the most up-to-date and comprehensive frameworks on the topic, as it considers emerging AI risks, addresses governance and response mechanisms comprehensively, and also includes an agentic AI risk management framework (do not miss Appendix 2).这是该领域目前最前沿、最全面的框架之一，因为它考虑了新兴的人工智能风险，全面探讨了治理和应对机制，还包含了一个智能体（Agentic AI）风险管理框架（千万不要错过附录 2）。
Read alongside other developments, version 3.0 of this framework signals that China wants to position itself as a global leader in AI safety (a strategic position the U.S. is also interested in holding).结合其他相关进展来看，该框架的 3.0 版本表明中国希望将自身定位为人工智能安全领域的全球领导者（这也是美国同样有意占据的战略地位）。
I will keep you posted. Download the framework and share it with your team.我会持续为您跟进。请下载该框架并与您的团队分享。
Are there binding regulations associated with this? I’m obviously not familiar with how China regulates businesses, but this framework (while incredibly detailed) doesn’t seem to have proscriptive requirements.
Whether a framework can be audited against is a separate question from whether it is well written. ISO/IEC 23894:2023, published in February 2023, is the closest international analogue on AI risk management, and ISO's own abstract calls it guidance whose application can be customised to any organisation and its context. That makes it useful as a checklist and thin as evidence for a customer or an insurer, and ISO has not said whether a requirements version is planned. A small company with no AI team gets the same trade from Framework 3.0: a structure to copy, and nothing to hand anyone afterwards. The agentic appendix is the part worth reading first.