Steven Sloman and Philip Fernbach · 2017
Why We Never Think Alone
Cover via Open LibraryRough AI truth score
75/100
The book's empirical foundation is strong: people routinely overestimate mechanistic understanding and rely on knowledge distributed across other people, tools, and institutions. The corrective is less automatic. Asking for explanations can reveal gaps, but awareness of dependence does not by itself identify trustworthy experts, repair misinformation, or improve collective decisions.
Based on three central claims · medium-high confidence
Repeated experiments find that people rate their understanding of devices and mechanisms highly, then reduce those ratings after attempting step-by-step explanations. The effect is strongest for explanatory knowledge and weaker for facts, procedures, and narratives, giving it clear scope rather than making it generic overconfidence.
Research on transactive memory, expertise attribution, tools, and information access shows that people remember who or where can supply knowledge and often confuse access with possession. Distribution creates real capability but also dependence, coordination costs, and vulnerability to unreliable sources.
Explanation prompts and calibrated humility can expose gaps and improve information seeking, but they do not guarantee accurate source selection, aggregation, or action. Expertise is hard to evaluate, networks can amplify common error, and institutions require incentives, verification, and accountability beyond individual humility.
What this number means. It is an AI-generated first-pass judgment of three central factual or causal claims—not a rating, exhaustive fact-check, or human peer review. Claim credits are 100% for supported, 75% for mostly supported, 50% for mixed, and 25% for weak, then averaged and rounded. Lower confidence means the score should move more as better evidence arrives.
Method three-central-claims/0.1.0 · checked 2026-09-01 · 3/3 selected claims assessed · method and source audit