The European Regulation on Artificial Intelligence, or E.U. AI Act (Act), approved in March 2024 as the world’s first comprehensive AI law, will have a global impact, applying to any companies whose AI interacts with Europeans. It promises regulation proportionate to the risks of each AI use, but companies worry that the Act’s layers of red tape will hold back innovation. This second article in a three-part series that answers common questions about the Act addresses the implications of the Act’s risk tiers and required assessments, including how it treats internal uses of AI. Part one reported on reactions among companies and delved into gray areas around its reach, the Act’s categorization of AI providers and deployers, and requirements affecting use of generative AI features. Part three will discuss practical steps for companies to prepare now for the Act. See “IBM, eBay and Walgreens CPOs Outline 10 Steps for Building AI Governance” (Oct. 18, 2023).