2024关键基础设施中人工智能的角色与责任框架报告

Roles and Responsibilities Frameworkfor Artificial Intelligence In Consultation withThe Artificial Intelligence Safety and Security Board November 14, 2024 Contents Artificial Intelligence Safety and Security Board Membership5 Executive Summary6 I. Introduction IV. Roles and Responsibilities for the Safe and Secure Deployment of AI in CriticalInfrastructure A. Overview12B. Key Terms and Structure12C.The Framework13I. Cloud and Compute Infrastructure Providers14 V. Conclusion A. Desired Outcomes of Framework 28 Appendix A: AI Roles and Responsibilities Matrix30 Appendix B: Glossary31 Appendix C: Notes33 Letter from the Secretary America’s critical infrastructure systems – systems that power our homes and businesses, deliverclean water, facilitate the digital networks that connect us, and much more – are vital to domesticand global safety and stability. Disruptions to the smooth operation of our nation’s hospitals andemergency services, electricity substations, pipelines, water treatment facilities, and other critical Artificial intelligence (AI) is already altering the way Americans interface with criticalinfrastructure. New technology, for example, is helping to sort and distribute mail to Americanhouseholds, quickly detect earthquakes and predict aftershocks, and prevent blackouts and otherelectric-service interruptions.These uses do not come without risk, though: a false alert of an Meanwhile, our adversaries are using and will continue to use AI to launch complex, sophisticated,and frequent cyberattacks on U.S. critical infrastructure.To counter these threats, criticalinfrastructure operators can use AI to both more effectively defend against malicious attacks and America’s continued security and prosperity will depend on how critical infrastructure The Department of Homeland Security (DHS) is charged with safeguarding America’s criticalinfrastructure in our evolving and expanding threat environment.To inform our approach to thisvital task, earlier this year I convened some of the nation’s foremost leaders in technology, industry,civil rights, academia, and government to form a first-of-its-kind AI Safety and Security Board. In This Framework proposes specific actions that entities across the AI ecosystem can consider,adapt, and implement based on their role and relationship to the critical services that ultimatelyserve the American people. It is designed to be incorporated into new and existing processes thatdrive product development, procurement decisions, and information exchanges. It also captures Addressing the extraordinary scale and impact of AI in U.S. critical infrastructure calls for a whole-of-nation approach.This Department has a critical role to play in socializing and harmonizing coreelements of this Framework, which draws from the important work of the White House, the AISafety Institute, and our Cybersecurity and Infrastructure Security Agency. Going forward, we willengage with our critical infrastructure partners to understand how this Framework can be adapted can harmonize our approach to AI safety and security across critical infrastructure globally. As the Secretary of Homeland Security and Chair of the AI Safety and Security Board, I am proud ofwhat we have achieved together. We welcome continued dialogue on these important issues, and Alejandro N. MayorkasSecretary, U.S. Department of Homeland Security Artificial Intelligence Safety and Security The Artificial Intelligence Safety and Security Board (the “Board”) advises the Secretary, the criticalinfrastructure community, other private sector stakeholders, and the broader public on the safe, secure,and responsible development and deployment of AI technology in our nation’s critical infrastructure.Theduties of the Board are solely advisory in nature. Board Members serve as uncompensated representativesof their sectors. The Board provided advice, information, and recommendations in the development of the Alejandro N. Mayorkas,Secretary,U.S. Department of Homeland Security; Chair, Artificial Intelligence Safetyand Security BoardSam Altman,CEO, OpenAIDario Amodei,CEO and Co-Founder, AnthropicEd Bastian,CEO, Delta Air LinesMarc Benioff,Chair and CEO, SalesforceRumman Chowdhury, Ph.D.,CEO, Humane IntelligenceMatt Garman,CEO, Amazon Web ServicesAlexandra Reeve Givens,President and CEO, Center for Democracy and TechnologyBruce Harrell,Mayor of Seattle, Washington; Chair,Technology and Innovation Committee, United StatesConference of MayorsDamon T. Hewitt,President and Executive Director, Lawyers’ Committee for Civil Rights Under LawVicki Hollub,President and CEO, Occidental PetroleumJensen Huang,President and CEO, NVIDIAArvind Krishna,Chairman and CEO, IBMFei-Fei Li, Ph.D.,Co-Director, Stanford Human-centered Artificial Intelligence InstituteWes Moore,Governor of MarylandSatya Nadella,Chairman and CEO, MicrosoftShantanu Narayen,Chair and CEO, AdobeSundar Pichai,CEO, Alphabet Executive Summary Artificial Intelli