AI Governance and Its Impact on Society: From Rules for Machines to Protection for People

AI governance is often discussed as rules for machines—risk tiers and compliance checklists—rather than protection for the people affected by automated decisions. Drawing on four influential instruments and six documented cases, this Viewpoint proposes five normative principles for judging governance from the affected individual’s standpoint: proportionate safeguards, meaningful oversight, actionable transparency, traceable lifecycle responsibility, and equitable distribution of benefits and harms. Its central claim is that the instruments largely agree on what to promise, yet failures persist because those promises are seldom turned into mechanisms for individuals to learn of, contest and obtain redress for automated decisions.