Regulators, news media and legislators are asking increasingly detailed questions about organizations’ artificial intelligence and machine learning (AI/ML) systems. Five financial regulatory agencies recently asked companies to answer a list of questions about their AI/ML use and the European Commission asked for comments on 80 pages of proposed AI regulations. This article, part two in a three-part series, provides questions that companies can ask themselves to guide oversight of their AI use before outsiders demand answers. With each question, we offer insight on steps to take and developments to watch. Part one
covered compliance essentials for AI/ML tools, while part three will cover how AI audits work and describe the Federal Reserve System’s recommended “three lines of defense” framework to manage algorithm risks. See “How to Achieve Trustworthy AI Using the European Commission’s Final Assessment List
” (Aug. 5, 2020).