Book

As artificial intelligence becomes embedded in government, finance, development institutions and corporate decision-making, a fundamental question emerges: when systems can generate recommendations, assessments and decisions at unprecedented speed, who remains responsible for their consequences? Automation promises greater efficiency, consistency and scale. Yet institutional legitimacy cannot be reduced to computational accuracy. Evidence must be examined, authority must be established, uncertainty must be acknowledged, and decisions must remain open to scrutiny, correction and challenge. Institutional Judgment Under Automation explores the critical distinction between automated intelligence and accountable institutional judgment. It examines how organisations can adopt increasingly sophisticated technologies without surrendering the human responsibility, procedural safeguards and institutional authority on which trustworthy decisions depend. Moving beyond conventional debates about artificial intelligence, the book presents governance as an essential part of technological implementation rather than an obstacle to innovation. It considers the relationships between evidence, decision rights, human oversight, institutional memory and demonstrable accountability. At its centre is a practical principle: technology may accelerate analysis, but it cannot independently confer legitimate authority. Written for policymakers, institutional leaders, researchers, development practitioners, governance professionals and those responsible for deploying AI, this book offers a framework for understanding what should be automated, what must remain subject to human judgment, and how both can operate within defensible institutional systems. The future of responsible automation will not be determined by technological capability alone. It will depend on whether institutions can preserve the integrity of judgment while transforming the machinery through which decisions are made. The challenge is not simply to build more intelligent systems. It is to ensure that institutions remain capable of making, explaining and defending the decisions those systems inform.