Three ways we collaborate with AI
AI Fluency is the ability to interact with AI in ways that are effective, efficient, ethical, and safe — a blend of practical skills, knowledge, and values, not just tool mechanics. The course frames three emerging modes of collaboration.
| Mode | What it is |
|---|---|
| Automation | The AI completes specific tasks based on your instructions. |
| Augmentation | You and the AI collaborate as thinking and execution partners. |
| Agency | You configure the AI to work independently on your behalf — setting its knowledge and behavior rather than giving it one task. |
Four competencies that work together
Developed by Prof. Rick Dakan and Prof. Joseph Feller, the 4D framework names the four competencies that, combined, make collaboration work.
Key concept · Delegation, Description, Discernment, Diligence
- Delegation — deciding what work goes to humans, what goes to AI, and how to split it; understanding your goals and the AI's capabilities to make strategic choices.
- Description — communicating with AI effectively: defining the outputs you want, guiding the process, and specifying how the AI should behave.
- Discernment — critically evaluating AI outputs, processes, and behaviors: assessing quality, accuracy, and appropriateness.
- Diligence — using AI responsibly and ethically: thoughtful choices, transparency, and accountability for AI-assisted work.
Saying what you want — three facets, six techniques
Description has three facets, and Discernment mirrors them one-to-one: you describe the Product (the output — format, style, length, detail), the Process (how to approach the task — methods, frameworks, steps), and the Performance (how the AI should behave with you — concise or detailed, challenging or supportive).
Key concept · Six prompting techniques (plus the secret weapon)
Prompt engineering is just designing effective instructions. Six foundations: give context (what, why, background); show examples of the output you want; specify constraints (format, length); break complex tasks into steps; ask the AI to think first; and define the AI's role or tone. The "secret weapon": ask the AI itself to help improve your prompt. Prompting is iterative — expect to refine based on results. (Claude 101 frames the same idea as set the stage → define the task → specify rules.)
Discernment, the loop, and responsibility
Discernment evaluates the same three facets you described: Product discernment (is the output good?), Process discernment (did it approach the task well?), and Performance discernment (was its collaborative behavior right?).
Key concept · The Description–Discernment loop
Real work is a cycle: describe what you need → discern the quality of what you get → refine your description → integrate your own expertise and judgment, taking responsibility for the final output. Repeat per task. This loop is where human–AI collaboration produces results that exceed what either could achieve alone.
Key concept · Diligence closes the loop
Where Delegation, Description, and Discernment chase effectiveness and efficiency, Diligence covers the ethical and safe side: making thoughtful choices about which systems to use, being transparent about AI's role in your work, and taking accountability for AI-assisted outputs. Fluency is all four D's together — drop Diligence and you're fast but not responsible.
Chapter 3 quiz
Fifteen questions on the modes, the 4Ds, Description, and the loop. Answer first, then expand Show answer.
Part A · Multiple choice
AI Fluency aims for interactions that are:
- A Effective, efficient, ethical, and safe
- B Fast and cheap
- C Fully automated
- D Code-free
Show answer
A. The four "E/E/E/S" goals.
In which collaboration mode do you configure the AI to work independently on your behalf?
- A Agency
- B Automation
- C Augmentation
- D Abstraction
Show answer
A — Agency. Automation = specific tasks; Augmentation = partnership.
Which competency decides what work goes to humans vs. AI?
- A Delegation
- B Description
- C Discernment
- D Diligence
Show answer
A — Delegation.
Which competency covers ethics, transparency, and accountability?
- A Diligence
- B Description
- C Delegation
- D Discernment
Show answer
A — Diligence.
The three facets of Description are:
- A Product, Process, Performance
- B Past, Present, Future
- C Plan, Prompt, Publish
- D People, Product, Price
Show answer
A. Discernment mirrors the same three.
Which is one of the six prompting techniques?
- A Show examples of the output you want
- B Always use one word
- C Never give context
- D Avoid specifying format
Show answer
A. Plus context, constraints, steps, think-first, and role/tone.
The "secret weapon" of prompting is:
- A Ask the AI to help improve your prompt
- B Use all caps
- C Send the same prompt repeatedly
- D Never iterate
Show answer
A.
Part B · True or false
Prompting is a one-shot skill — a good prompt gets it right the first time.
Show answer
False. It's iterative (and collaborative with the AI); expect to refine.
Discernment evaluates the product, the process, and the AI's performance.
Show answer
True. It mirrors the three facets of Description.
In the Description–Discernment loop, you take responsibility for the final output.
Show answer
True. You integrate your own judgment and own the result.
Performance description is about the format of the output file.
Show answer
False. That's Product description. Performance is how the AI behaves with you (concise/detailed, challenging/supportive).
Part C · Short answer
Name the four competencies of the 4D framework.
Show answer
Delegation, Description, Discernment, and Diligence.
Name the three modes of human–AI collaboration.
Show answer
Automation, Augmentation, and Agency.
Name four of the six prompting techniques.
Show answer
Any four: give context, show examples, specify constraints, break into steps, ask it to think first, define role/tone.
What are the four steps of the Description–Discernment loop?
Show answer
Describe, discern, refine, and integrate your own expertise (owning the final output).