Trust After Thinking Machines: Silent Authority, Human Responsibility, and the Future of Legitimate Power

Trust After Thinking Machines follows a simple observation to its uncomfortable end: once institutions can 'think' at scale, they stop treating intelligence as a scarce human resource and start treating judgment as a renewable utility. Automation arrives quietly—as triage screens, risk scores, routing queues, ranking systems, and policy engines—reshaping who gets served, flagged, priced, or denied. The book argues this is not merely a technical shift but a shift in authority itself. The core crisis is not that automated decisions are sometimes wrong; it is that they become unanswerable. Responsibility diffuses across vendors, models, policies, and committees until no one can say with evidence who decided, why, and who owns the outcome. The book builds a practical theory of accountable authority for the agentic era, proposing three enforceable properties for legitimate machine-mediated judgment: boundedness, meaningful contestability, and identifiable responsibility—translated into infrastructure like decision provenance records, reviewable evidence trails, and disagreement architectures.