AI generates inputs, you synthesize judgment — outsourcing…

When using AI systems, maintain explicit separation between input generation (AI-assisted) and judgment synthesis (human-retained), because outsourcing the integration function constitutes epistemic abdication regardless of input quality.

When This Fires

When designing any AI-assisted workflow
When you notice yourself accepting AI conclusions without performing the synthesis step
When AI output includes integrated judgments ("you should do X") rather than inputs ("here are the options with trade-offs")
Complements Apply the defense test to AI conclusions (defense test for AI conclusions) with the structural workflow design

Common Failure Mode

Asking AI for conclusions instead of inputs: "What should I do about [complex situation]?" This outsources the synthesis. Better: "What are the key factors, options, and trade-offs in [complex situation]?" Then you synthesize. The AI provides richer inputs than you could gather alone; you provide the judgment that makes the decision yours.

The Protocol

(1) When using AI, structure your requests to generate inputs, not conclusions. Ask for: options and trade-offs, relevant data, multiple perspectives, risk assessments, historical precedents. Avoid asking for: recommendations, decisions, "what should I do?" (2) Receive the AI-generated inputs. (3) Perform the synthesis yourself: weigh the inputs against your values, context, and risk tolerance. Consider what the AI might be missing about your specific situation. Reach a conclusion you can defend independently (Apply the defense test to AI conclusions). (4) The conclusion is yours — informed by AI inputs but synthesized through your judgment. (5) If you catch yourself accepting AI conclusions without synthesis → redesign the prompt to request inputs instead.