AI-CI safety guidelines
governmentAn official U.S. government document issued under EO 14110; useful for understanding how federal agencies are translating AI safety principles into actionable guidelines for high-stakes infrastructure operators.
Metadata
Importance: 62/100guidance documentprimary source
Summary
The Department of Homeland Security's April 2024 guidelines provide a four-part framework (Govern, Map, Measure, Manage) to help critical infrastructure owners and operators manage AI-related risks. The document identifies three cross-sector risk categories—attacks using AI, attacks on AI systems, and AI design/implementation failures—and maps mitigation strategies to the NIST AI Risk Management Framework. It represents a practical, sector-agnostic approach to operationalizing AI risk management in high-stakes infrastructure contexts.
Key Points
- •Establishes a four-part AI risk management framework: Govern, Map, Measure, and Manage, aligned with the NIST AI RMF.
- •Identifies three critical AI risk categories: adversarial use of AI, attacks targeting AI systems, and failures from AI design or implementation flaws.
- •Mandated by Executive Order 14110, targeting critical infrastructure sectors such as energy, water, transportation, and financial services.
- •Provides concrete mitigation strategies for each risk category, including general mitigations applicable across sectors.
- •Emphasizes organizational culture and governance as foundational to effective AI risk management, not just technical controls.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| Cyberweapons Risk | Risk | 91.0 |
Cached Content Preview
HTTP 200Fetched Mar 15, 202678 KB
Publication: April 2024
Department of Homeland Security
MITIGATING ARTIFICIAL INTELLIGENCE (AI) RISK:
Safety and Security Guidelines
for Critical Infrastructure Owners
and Operators
-- 1 of 28 --
2
Table of Contents
EXECUTIVE SUMMARY ................................................................................................................................................. 4
INTRODUCTION ............................................................................................................................................................ 5
AI RISKS TO CRITICAL INFRASTRUCTURE................................................................................................................... 7
AI Uses and Patterns of Adoption .......................................................................................................................... 7
Cross-Sector AI Risk Categories ............................................................................................................................. 9
GUIDELINES FOR CRITICAL INFRASTRUCTURE OWNERS AND OPERATORS .......................................................... 10
Govern: Establish an organizational culture of AI risk management. ................................................................ 11
Map: Understand your individual AI use context and risk profile. ...................................................................... 11
Measure: Develop systems to assess, analyze, and track AI risks. ................................................................... 12
Manage: Prioritize and act upon AI risks to safety and security. ....................................................................... 13
CONCLUSION.............................................................................................................................................................. 15
APPENDIX A: CROSS-SECTOR AI RISKS AND MITIGATION STRATEGIES ................................................................. 16
Risk Category: Attacks Using AI ............................................................................................................................ 16
Risk Category: Attacks on AI ................................................................................................................................. 17
Risk Category: AI Design and Implementation Failures ...................................................................................... 19
General Mitigations for AI Risks ........................................................................................................................... 20
APPENDIX B: GUIDELINES MAPPED TO NIST AI RMF .............................................................................................. 22
Govern ................................................................................................................................................................... 22
Map ......................................................
... (truncated, 78 KB total)Resource ID:
7786ae9986ce7a71 | Stable ID: sid_AXgbWwyBwr