AI can compare options, predict outcomes, and produce a persuasive recommendation in seconds. But when that recommendation changes a life, who has actually made the decision-and who remains answerable for it? In medicine, employment, education, law, public services, safety, and private life, artificial intelligence is moving from supplying information to shaping action. The danger is not simply that a system may be wrong. It is that responsibility can disappear inside an efficient process: the model recommends, the interface directs, the human approves, and no one can explain why the outcome was justified. In When Recommendation Becomes Responsibility, G B Edlin examines the moral distance between advice and decision. He explores automation bias, hidden values, the moral crumple zone, meaningful human oversight, practical wisdom, dissent, contestability, and the institutional conditions that allow conscience to function. Clear, balanced, and written for general readers, this substantial third volume of The Borrowed Mind includes: - five moral lenses for AI-assisted choices;- practical tests for meaningful human oversight;- a full worked responsibility audit;- decision-rights and stop-rule frameworks;- a maturity model for accountable organisations;- case studies and practice scenarios across high-stakes fields; and>This is neither an argument against AI nor a defence of human judgment as it currently exists. It is a constructive guide to building combined systems that use machine capability without surrendering human agency, moral attention, or the duty to repair harm. The machine can recommend. The decision still needs an author.
