Artificial intelligence is changing more than how institutions process information. It is changing how they decide.As AI systems increasingly detect patterns, rank risks, recommend actions, and operate across organizational boundaries, a critical question emerges: Who is actually responsible for the judgment?The Institution in the Loop examines what happens when intelligence is distributed across people, algorithms, agencies, companies, policies, and technical systems, while authority and accountability remain fragmented.Joe B. Williams, PhD, develops a practical framework for institutional judgment in the age of AI. Drawing on public government reports, safety investigations, AI governance standards, and research, the book explores how institutions can use increasingly capable AI without surrendering the responsibilities that make consequential decisions legitimate.Readers will examine how institutions can: distinguish signals from conclusionscalibrate confidence, consequence, and reversibilityprevent human oversight from becoming a ceremonial checkboxdefine who may recommend, authorize, execute, override, and review AI-enabled decisionspreserve meaningful dissentgovern information and judgment across organizational boundariesrecognize adversarial manipulation and deceptionlearn from errors, near misses, and unexpected consequencesmaintain accountability as AI systems become more autonomous and capableThe central argument moves beyond the familiar demand to keep a "human in the loop."A human presence alone is not enough.The institution itself must remain in the loop.AI can extend an institution's mind. It cannot become its conscience.
