《麻省理工学院（MIT）：2024年AI风险存储库报告：人工智能风险的综合元评审、数据库和分类法（英文版）（79页）.pdf》由会员分享，可在线阅读，更多相关《麻省理工学院（MIT）：2024年AI风险存储库报告：人工智能风险的综合元评审、数据库和分类法（英文版）（79页）.pdf（79页珍藏版）》请在三个皮匠报告上搜索。

1、The AI Risk Repository:A ComprehensiveMeta-Review,Database,and Taxonomy ofRisks From Artificial IntelligencePeter Slattery1,2,Alexander K.Saeri1,2,Emily A.C.Grundy1,2,Jess Graham3,Michael Noetel2,3,Risto Uuk4,5,James Dao6,Soroush Pour6,Stephen Casper7,and Neil Thompson1.1MIT FutureTech,Massachusetts

2、 Institute of Technology,2Ready Research,3School of Psychology,The University ofQueensland,4Future of Life Institute,5KU Leuven,6Harmony Intelligence,7Computer Science and Artificial IntelligenceLaboratory,Massachusetts Institute of Technology.Correspondence to pslatmit.edu.AbstractThe risks posed b

3、y Artificial Intelligence(AI)are of considerable concern to academics,auditors,policymakers,AI companies,and the public.However,a lack of shared understanding of AI riskscan impede our ability to comprehensively discuss,research,and react to them.This paperaddresses this gap by creating an AI Risk R

4、epository to serve as a common frame of reference.This comprises a living database of 777 risks extracted from 43 taxonomies,which can be filteredbased on two overarching taxonomies and easily accessed,modified,and updated via our websiteand online spreadsheets.We construct our Repository with a sys

5、tematic review of taxonomies andother structured classifications of AI risk followed by an expert consultation.We develop ourtaxonomies of AI risk using a best-fit framework synthesis.Our high-level Causal Taxonomy of AIRisks classifies each risk by its causal factors(1)Entity:Human,AI;(2)Intentiona

6、lity:Intentional,Unintentional;and(3)Timing:Pre-deployment;Post-deployment.Our mid-level Domain Taxonomyof AI Risks classifies risks into seven AI risk domains:(1)Discrimination&toxicity,(2)Privacy&security,(3)Misinformation,(4)Malicious actors&misuse,(5)Human-computer interaction,(6)Socioeconomic&e