Skip to main contentSkip to footer
Try Claude

AI Fluency:
The Description-Discernment Loop

Learn to collaborate with AI systems effectively, efficiently, ethically, and safely

12 modules • 3-4 hours

Estimated Time: 30 - 60 minutes

What you’ll learn

By the end of this lesson, you’ll be able to:

  • Apply Description and Discernment skills to a real project
  • Engage in productive Description-Discernment feedback loops
  • Create results through human-AI collaboration that exceed what either could achieve alone

Exercises

Exercise: Project Execution with Description-Discernment Loops

Estimated time: 30-60 minutes

Now it's time to put everything you've learned into practice by working on the project you planned in Lesson 5, using the Description and Discernment skills you've been developing.

Step 1: Review Your Project Plan

  • Pull up the project plan you created in Lesson 5
  • Quickly review your delegation decisions about which tasks would benefit from human expertise, AI capabilities, or collaboration
  • Feel free to refine your plan based on what you've learned since then

Step 2: Prepare Your Description Approach

Start a conversation with Claude and explain the project you'll be working on together. Before diving into execution, plan how you'll approach Description:

  • Product Description: What specific outputs do you need from Claude for each task? What format, style, length, and level of detail are you looking for?
  • Process Description: How should Claude approach each task? Are there specific methods, frameworks, or steps you want it to follow?
  • Performance Description: What kind of collaborative behavior do you want from Claude during this project? Should it be concise or detailed, challenging or supportive, focused on ideas or analysis?

Discuss these questions with Claude to establish clear expectations for your collaboration.

Step 3: Execute Your Project Using Description-Discernment Loops

Now, work through your planned project tasks with Claude. For each task:

  1. Describe what you need clearly, using the Description skills you've learned:
    • Be specific about what you want (Product)
    • Guide how Claude should approach or think about the task (Process)
    • Specify how you want Claude to engage with you during the process (Performance)
  2. Discern the quality of what you receive:
    • Evaluate the output itself (Product Discernment)
    • Assess how Claude approached the task (Process Discernment)
    • Consider if Claude's behavior is most helpful for what you need (Performance Discernment)
  3. Refine based on your discernment:
    • Provide feedback on what worked and what didn't
    • Clarify or adjust your description as needed
    • Request iterations until you're satisfied with the result
  4. Integrate your own expertise and judgment:
    • Add your unique perspective, creativity, or domain knowledge
    • Make the final decisions about what to keep, modify, or discard
    • Take responsibility for the final output

Continue this Description-Discernment loop for each task in your project until completion.

Reflection

Before moving on, take a moment to consider:

  • What patterns did you notice in the types of descriptions that led to the best outcomes?
  • Which required more effort from you: Description or Discernment? Why do you think that was the case?
  • How did your actual project execution compare to your initial plan from Lesson 5? What adjustments did you make along the way?
[@portabletext/react] Unknown block type "horizontalRule", specify a component for it in the `components.types` prop

What’s next

In the next lesson, we'll explore the final competency in the AI Fluency Framework: Diligence. While Delegation, Description, and Discernment focus primarily on effectiveness and efficiency, Diligence addresses the ethical and safety aspects of working with AI. You'll learn how to ensure your AI collaborations are responsible, transparent, and accountable.

Feedback on this course

As you progress through the course, we'd love to hear from you about how you are using concepts from the course in your life, work, or classes and any feedback you may have. Share your feedback here.

Acknowledgments and license

Copyright 2025 Rick Dakan, Joseph Feller, and Anthropic. Released under the CC BY-NC-SA 4.0 license.

This course is based on The AI Fluency Framework by Dakan and Feller.

Supported in part by the Higher Education Authority, Ireland, through the National Forum for the Enhancement of Teaching and Learning.