Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Emergency Response Measures for Catastrophic AI Risk

James Zhang, Miles Kodama, Zongze Wu, Michael Chen, Yue Zhu, Geng Hong

TL;DR

The paper addresses catastrophic AI risk in China and proposes Frontier Safety Policies (FSPs) as a practical mechanism to operationalize the preventive and warning phases of the national emergency framework. It analyzes existing Chinese regulations (Interim Measures, TC260 governance, emergency guidelines, and GB/T 45654-2025) and self-governance efforts (Shanghai Frontier Risk Framework, industry commitments) to show how FSPs could be embedded in policy and industry practice. It situates these ideas in an international context, drawing from the EU AI Act, US state approaches (California SB53, NY RAISE), and industry commitments (AI Seoul Summit) to illustrate a common trend toward capability-based risk tiers and rapid reporting. The paper argues that formal adoption of FSPs could strengthen AI emergency preparedness across prevention, surveillance, response, and recovery, while noting limitations and suggesting concrete regulatory extensions and standards alignment.

Abstract

Chinese authorities are extending the country's four-phase emergency response framework (prevent, warn, respond, and recover) to address risks from advanced artificial intelligence (AI). Concrete mechanisms for the proactive prevention and warning phases, however, remain under development. This paper analyzes an implementation model inspired by international AI safety practices: frontier safety policies (FSPs). These policies feature pre-deployment evaluations for dangerous capabilities and tiered, pre-planned safety measures. We observe close alignment between FSPs and the proactive phases of China's emergency response framework, suggesting that the FSP model could help operationalize AI emergency preparedness in a manner consistent with China's established governance principles.

Emergency Response Measures for Catastrophic AI Risk

TL;DR

The paper addresses catastrophic AI risk in China and proposes Frontier Safety Policies (FSPs) as a practical mechanism to operationalize the preventive and warning phases of the national emergency framework. It analyzes existing Chinese regulations (Interim Measures, TC260 governance, emergency guidelines, and GB/T 45654-2025) and self-governance efforts (Shanghai Frontier Risk Framework, industry commitments) to show how FSPs could be embedded in policy and industry practice. It situates these ideas in an international context, drawing from the EU AI Act, US state approaches (California SB53, NY RAISE), and industry commitments (AI Seoul Summit) to illustrate a common trend toward capability-based risk tiers and rapid reporting. The paper argues that formal adoption of FSPs could strengthen AI emergency preparedness across prevention, surveillance, response, and recovery, while noting limitations and suggesting concrete regulatory extensions and standards alignment.

Abstract

Chinese authorities are extending the country's four-phase emergency response framework (prevent, warn, respond, and recover) to address risks from advanced artificial intelligence (AI). Concrete mechanisms for the proactive prevention and warning phases, however, remain under development. This paper analyzes an implementation model inspired by international AI safety practices: frontier safety policies (FSPs). These policies feature pre-deployment evaluations for dangerous capabilities and tiered, pre-planned safety measures. We observe close alignment between FSPs and the proactive phases of China's emergency response framework, suggesting that the FSP model could help operationalize AI emergency preparedness in a manner consistent with China's established governance principles.

Paper Structure

This paper contains 11 sections, 2 figures, 1 table.

Table of Contents

  1. Motivating Emergency Response Measures
  2. Regulating Frontier AI in China
  3. Government Regulatory Frameworks
  4. Self-Governance Initiatives
  5. International Approaches to AI Emergency Response
  6. The European Union AI Act: General-Purpose AI Code of Practice
  7. State-level Emergency Preparedness in the United States
  8. Industry Best Practices and International Commitments
  9. Key Considerations and Future Directions
  10. Acknowledgments
  11. Implementations for the four phases of emergency response

Figures (2)

  • Figure 1: Definition and key elements of a frontier safety policy.
  • Figure 2: Registry-centric model lifecycle illustrating China’s current regulatory ecosystem alongside our proposed additions: requiring providers to develop Frontier Safety Policies (FSPs) and integrating a catastrophic-risk test suite into pre-deployment assessments.