Course Challenge
Bring all four competencies together: a scenario-based knowledge check plus a capstone where you take a real task through Delegation, Description, Discernment, and Diligence.
How the Challenge Works
The challenge has two parts:
- Knowledge check: ten scenario questions covering every module. Aim for at least 8 out of 10. If you miss one, read the explanation and revisit that lesson.
- Capstone task: a hands-on exercise that takes a real piece of your own work through all four competencies.
This course is free and self-paced, so there’s nothing to submit. The goal is to leave with a way of working with AI that you can apply to tools that don’t exist yet.
Part 1: Knowledge Check
1. (Module 1) A coworker says, 'I'm already good at AI. I have a folder of 50 prompt templates.' What's missing from that idea of fluency?
Fluency covers all four competencies. Templates help with Description but don't tell you whether to use AI, whether the output is right, or whether you're using it responsibly.
2. (Module 1) You configure an AI assistant to automatically draft replies to every incoming support ticket overnight. Which mode is this, and what does it demand most?
AI acting on your behalf across future situations is agency. Its mistakes repeat, so the setup and monitoring carry the most weight.
3. (Module 2) A manager plans to have AI decide which three team members get promoted, based on their project lists. What's the core delegation problem?
Problem awareness: promotions are a judgment the manager must own. AI might help organize evidence, but delegating the decision hands off the part that matters most.
4. (Module 2) Which task best fits 'automate, then spot-check'?
High fit, easy to check, and low stakes: the ideal automation profile.
5. (Module 3) Your prompt asks for 'a report on our hiring process,' and the AI invents statistics. Which description fix helps most?
Describing the process (which sources to use) and boundaries (placeholders instead of guesses) prevents fabricated figures.
6. (Module 3) You're practicing for a job interview and want AI to help. Which is a performance description?
Performance description shapes how the AI behaves as a collaborator: its role, pacing, and how direct its feedback is.
7. (Module 4) An AI summary of a 30-page policy is clear and well organized. What should you check first?
Product discernment starts with accuracy and completeness, checked against the source. Polish and confidence aren't evidence.
8. (Module 4) After several rounds of feedback, the AI's drafts keep getting worse and the chat is cluttered with failed attempts. What's the best move?
When a conversation is anchored on a bad direction, restarting with an improved description, informed by your discernment, is usually faster.
9. (Module 5) A student uses AI to brainstorm essay topics, which their instructor's policy allows with disclosure. What does transparency diligence require?
Transparency means following the stated policy and being honest with the people who have a stake, here the instructor.
10. (Module 5) A team adds 'AI-generated' to a client report but doesn't check its figures. One is wrong. Who's responsible?
Deployment diligence: you own what you share. Transparency and verification are separate duties, and disclosure doesn't excuse errors.
Part 2: Capstone Task
Choose one real task you need to complete in the next few weeks, something with enough substance that you’d care if it went wrong. Work through it with an AI assistant, documenting each stage in a short write-up (bullet points are fine).
- Delegation (Module 2).
- Write the real goal and what success looks like.
- Break the task into pieces and mark each You, Augment, or Automate, using the five-question delegation check.
- Confirm the tool you’re using is allowed for the data involved.
- Description (Module 3).
- For the pieces AI will help with, write a description covering product, process, and performance.
- Use at least one technique: an example, a boundary, placeholders, or letting the AI interview you.
- Discernment (Module 4).
- Run the description–discernment loop at least three rounds. Record what you noticed, what caused it, and the feedback you gave.
- Apply your personal discernment checklist to the final output. Verify every specific fact against a source outside the AI.
- Diligence (Module 5).
- Decide what disclosure, if any, is appropriate, and why.
- Confirm you’d put your name on the result.
- Check that the work is consistent with your personal AI use statement. Update the statement if you learned something.
Self-assessment checklist. You’ve completed the capstone when you can say yes to all of these:
- I kept at least one piece of this work for myself, on purpose, and can explain why.
- My description covered product, process, and performance.
- I can point to at least two places where my discernment changed the result for the better.
- I verified every factual claim against a source outside the AI.
- I didn’t share anything I shouldn’t have, and I handled disclosure deliberately.
- I’d confidently put my name on the final result.
Congratulations, You’ve Finished AI Fluency: Foundations
You can now:
- Explain what AI fluency is and why it outlasts any single tool or trick
- Recognize automation, augmentation, and agency, and choose the right mode
- Delegate deliberately, piece by piece, based on your goal and your tools
- Describe the product, process, and performance you need
- Discern the quality of AI output, reasoning, and behavior, and turn that into better descriptions
- Practice diligence in how you create, disclose, and deploy AI-assisted work
Where to Go Next
- Keep the routine going. Use the 4D routine on one meaningful task each week.
- Share your AI use statement with your team or instructor and invite theirs.
- Start here if you skipped it: AI Fundamentals covers how AI works and the CRAFT prompting method in more depth.
- Go further with AI: AI & LLM Skills shows how language models work under the hood and how to start building with an AI API.
The 4D AI Fluency Framework was created by Rick Dakan and Joseph Feller with Anthropic and is released under CC BY-NC-SA 4.0. See aifluencyframework.org.