United States: Artificial Intelligence Risk Management Framework (AI RMF 1.0)

1. Part 1: Foundational Information 4
2. 1 Framing Risk 4
3. 1.1 Understanding and Addressing Risks, Impacts, and Harms 4
4. 1.2 Challenges for AI Risk Management 5
5. 1.2.1 Risk Measurement 5
6. 1.2.2 Risk Tolerance 7
7. 1.2.3 Risk Prioritization 7
8. 1.2.4 Organizational Integration and Management of Risk 8
9. 2 Audience 9
10. 3 AI Risks and Trustworthiness 12
11. 3.1 Valid and Reliable 13
12. 3.2 Safe 14
13. 3.3 Secure and Resilient 15
14. 3.4 Accountable and Transparent 15
15. 3.5 Explainable and Interpretable 16
16. 3.6 Privacy-Enhanced 17
17. 3.7 Fair – with Harmful Bias Managed 17
18. 4 Effectiveness of the AI RMF 19
19. Part 2: Core and Profiles 20
20. 5 AI RMF Core 20
21. 5.1 Govern 21
22. 5.2 Map 24
23. 5.3 Measure 28
24. 5.4 Manage 31
25. 6 AI RMF Profiles 33
26. Appendix A: Descriptions of AI Actor Tasks from Figures 2 and 3 35
27. Appendix B: How AI Risks Differ from Traditional Software Risks 38
28. Appendix C: AI Risk Management and Human-AI Interaction 40
29. Appendix D: Attributes of the AI RMF 42
30. Part 2 comprises the “Core” of the Framework. It describes four specific functions to help
31. Part 1: Foundational Information
32. 1. Framing Risk
33. 1.1 Understanding and Addressing Risks, Impacts, and Harms
34. 1.2 Challenges for AI Risk Management
35. 1.2.1 Risk Measurement
36. 1.2.2 Risk Tolerance
37. 1.2.3 Risk Prioritization
38. 1.2.4 Organizational Integration and Management of Risk
39. 2. Audience
40. 3. AI Risks and Trustworthiness
41. 3.1 Valid and Reliable
42. 3.2 Safe
43. 3.3 Secure and Resilient
44. 3.4 Accountable and Transparent
45. 3.5 Explainable and Interpretable
46. 3.6 Privacy-Enhanced
47. 3.7 Fair – with Harmful Bias Managed
48. 4. Effectiveness of the AI RMF
49. Part 2: Core and Profiles
50. 5. AI RMF Core
51. 5.1 Govern
52. 5.2 Map
53. 5.3 Measure
54. 5.4 Manage
55. 6. AI RMF Profiles
56. Appendix A:: Descriptions of AI Actor Tasks from Figures 2 and 3
57. Appendix B:: How AI Risks Differ from Traditional Software Risks
58. Appendix C:: AI Risk Management and Human-AI Interaction
59. 1. Human roles and responsibilities in decision making and overseeing AI systems
60. 2. Decisions that go into the design, development, deployment, evaluation, and use
61. 3. Human-AI interaction results vary. Under certain conditions – for example, in
62. 4. Presenting AI system information to humans is complex. Humans perceive and
63. Appendix D:: Attributes of the AI RMF
64. 2. Be consensus-driven and developed and regularly updated through an open, trans-
65. 3. Use clear and plain language that is understandable by a broad audience, including
66. 4. Provide common language and understanding to manage AI risks. The AI RMF
67. 5. Be easily usable and fit well with other aspects of risk management. Use of the
68. 6. Be useful to a wide range of perspectives, sectors, and technology domains. The AI
69. 7. Be outcome-focused and non-prescriptive. The Framework should provide a catalog
70. 8. Take advantage of and foster greater awareness of existing standards, guidelines, best
71. 9. Be law- and regulation-agnostic. The Framework should support organizations’
72. 10. Be a living document. The AI RMF should be readily updated as technology, under-

The outline comes from headings detected in the text, so it might miss some or include extras.