Skip to main content
← ExitWorking AI Fluency in Practice
0 / 6 lessons0 XP
1 / 6

In this lesson

Common workflows: draft → review → refine

Apply the draft → review → refine workflow to delegate mechanical subtasks to an AI while reserving critical judgment and synthesis for human review, consistent with Mollick and Mollick's AI-as-TOOL pedagogical frame…

You'll be able to

  • Apply the draft → review → refine workflow to delegate mechanical subtasks to an AI while reserving critical judgment and synthesis for human review, consistent with Mollick and Mollick's AI-as-TOOL pedagogical frame [^1].
  • Evaluate AI-generated first drafts against domain-specific quality criteria, identifying where the output requires correction, evidence strengthening, or structural revision before the artifact is suitable for professional use [^1][^6].
  • Create iterative refinement cycles in which you use AI critique (Mollick's AI-as-COACH role) to surface weaknesses in a draft, then apply your own judgment to accept, reject, or adapt suggested improvements [^1].
  • Classify which stages of a multi-step workflow benefit from AI automation (draft generation, scenario enumeration, initial classification) versus which demand human oversight (final approval, clinical safety review, alignment with regulatory standards), drawing on real-world clinician-in-the-loop and policy-development patterns [^2][^4][^5].
  • Explain why rewriting and debugging are more feasible when the AI produces a manipulable first draft than when starting from scratch, citing Papert's observation that easy revision transforms reluctance into iterative improvement [^6][^8].