AI FLUENCY · PHASE 2: APPRECIATION
Module 10 — The 4Ds of AI Fluency
What Is AI Delegation, and How Do We Decide Who Should Do What?
The first D of AI Fluency: deciding how work should be shared between humans and AI.
Who Should Do What?
Introduction: Should We Delegate Everything to AI?
AI can generate ideas, explain concepts, analyze information, draft documents, and perform selected tasks.
But just because AI can assist with an activity does not mean the entire activity should be assigned to an AI system.
Some tasks benefit from automation. Others benefit from collaboration. Some require direct human judgment and responsibility.
Before asking AI to do something, we should ask: Who should do what?
This is the central question of Delegation, the first D of AI Fluency.
Effective delegation begins by understanding the activity, recognizing AI capabilities, and deciding how responsibilities should be distributed.
1. What Is AI Delegation?
AI Delegation means thoughtfully deciding which activities should be performed by people, which can be supported by AI, and which may appropriately be executed by an AI system.
The objective is not to maximize the number of tasks assigned to AI.
The objective is to choose a suitable division of work that helps achieve a meaningful human goal.
Human Goal
↓
Understand the Activity
↓
Recognize AI Capabilities
↓
Decide Who Should Do What
An activity may involve different kinds of work, so effective delegation often requires examining individual tasks rather than treating the activity as one indivisible assignment.
2. Problem Awareness — Understand the Activity
The first dimension of effective delegation is Problem Awareness.
Before selecting an AI tool, we must understand what needs to be accomplished.
Questions to Consider
- What is the actual problem?
- What outcome do we want?
- Which activities are involved?
- Which decisions require human judgment?
- What are the consequences of errors?
Everyday Example
A teacher wants to help students understand the water cycle.
The actual goal is not merely to produce a document.
It is to improve students' understanding of evaporation, condensation, and precipitation.
Recognizing this goal helps the teacher decide which activities AI might support.
What needs to be accomplished?
3. Platform Awareness — Understand AI Capabilities
The second dimension is Platform Awareness.
Different AI systems have different capabilities, limitations, interfaces, and access to tools.
A general-purpose chatbot might generate explanations.
Another system might analyze uploaded documents.
A suitably connected system might perform selected actions through authorized applications.
Questions to Consider
- What can this particular AI system do?
- What information can it access?
- Which tools or applications can it use?
- How reliable is it for this task?
- What are its limitations?
- What permissions and safeguards are needed?
Technical capability does not automatically establish task suitability.
Even when an AI system can generate an answer, its output may still require verification, correction, or rejection.
What can this AI system appropriately do?
4. Task Delegation — Decide Who Should Do What
After understanding the problem and recognizing AI capabilities, we can decide how to distribute the activities involved.
Three categories are useful for understanding possible human–AI contributions.
👤
Human-Led
People establish goals, make important decisions, apply judgment, and retain accountability.
Example: Determining learning objectives.
🤝
AI-Assisted
AI offers explanations, ideas, analysis, or drafts for human evaluation and refinement.
Example: Suggesting teaching examples.
⚙️
AI-Executed
Suitable activities are performed automatically, with appropriate permissions, controls, and oversight.
Example: Automated document formatting.
These categories are not always mutually exclusive.
A single activity may contain human-led decisions, AI-assisted work, and automated execution.
The appropriate combination depends on the task, the available technology, and the consequences of possible errors.
5. Everyday Example: Preparing a Science Learning Guide
Imagine a teacher preparing a learning guide about the water cycle for school students.
The teacher could use the Delegation framework to decide how different activities should be handled.
Human-Led Activities
- Identify learning objectives.
- Understand students' needs.
- Determine suitable explanations.
- Verify scientific accuracy.
- Approve the final guide.
AI-Assisted Activities
- Suggest explanations.
- Generate everyday examples.
- Propose practice questions.
- Offer alternative lesson structures.
AI-Executed or Automated Activities
- Apply a defined document template.
- Perform selected formatting tasks.
- Generate approved file outputs when suitable tools are available.
The teacher remains responsible for reviewing the educational quality and suitability of the final material.
The goal is not maximum delegation. It is appropriate delegation that improves the learning activity.
6. The AI Delegation Framework
The three dimensions can be summarized in a simple framework:
STEP 1
PROBLEM AWARENESS
Understand the task and intended outcome.
↓
STEP 2
PLATFORM AWARENESS
Recognize AI capabilities and limitations.
↓
STEP 3
TASK DELEGATION
Choose appropriate roles for people and AI.
The framework is not a rigid one-time procedure.
As work progresses, new information may reveal that a task needs more human attention or that an AI system is unsuitable.
Effective delegation includes the willingness to reconsider earlier decisions.
7. When Should We Avoid Delegating to AI?
Not every activity benefits from AI delegation.
Delegation should be reconsidered when:
- The AI system cannot perform the activity reliably.
- Important information is unavailable or unsuitable for sharing with the system.
- The task involves unacceptable privacy, security, or safety risks.
- The system lacks necessary permissions or appropriate controls.
- Human expertise, personal understanding, or direct responsibility is essential.
- The activity cannot be supervised adequately.
Tasks with significant consequences may require stronger human review, specialist expertise, and formal safeguards.
Sometimes the most appropriate delegation decision is not to use AI.
AI Delegation at a Glance
| Dimension | Human-Led | AI-Assisted | AI-Executed |
|---|---|---|---|
| Main role | Human direction and judgment | AI support for human activity | Automated performance of selected tasks |
| Example | Define learning objectives | Suggest examples | Format a document using an approved template |
| Human involvement | Direct | Active collaboration | Oversight and exception handling as appropriate |
| Important consideration | Expertise and accountability | Critical evaluation | Reliability, permissions, and safeguards |
The categories describe different ways work may be distributed.
They are not universal rules that automatically determine the right approach for every situation.
Key Takeaway
Problem Awareness
Understand what needs to be accomplished.
Platform Awareness
Recognize what AI can appropriately contribute.
Task Delegation
Decide who should do what.
AI Delegation begins with human purpose, not with the AI tool.
Quick Knowledge Check
Try answering each question before revealing its explanation.
Question 1
What is the central question of AI Delegation?
A. How fast can AI generate text?
B. Who should do what?
C. How many AI tools can be used?
Reveal Answer
B. Delegation involves determining appropriate roles for humans and AI.
Question 2
What should come first when considering AI delegation?
A. Choosing the most advanced AI tool
B. Understanding the actual problem
C. Automating every available task
Reveal Answer
B. Problem Awareness begins by understanding the activity and desired outcome.
Question 3
What does Platform Awareness involve?
A. Recognizing AI capabilities and limitations
B. Assuming all AI tools are identical
C. Ignoring technical constraints
Reveal Answer
A. Platform Awareness means understanding what a particular system can and cannot appropriately do.
Question 4
Which example best represents an AI-assisted task?
A. A teacher deciding the learning goals
B. AI suggesting explanations for teacher review
C. Software automatically applying a fixed template
Reveal Answer
B. AI provides suggestions while the teacher remains actively involved.
Question 5
Does effective delegation mean assigning as many tasks as possible to AI?
A. Yes
B. No, it means choosing the appropriate division of work
C. Only for classroom activities
Reveal Answer
B. The goal is suitable delegation that supports the intended human outcome.
Question 6
When might it be appropriate not to delegate a task to AI?
A. When the system is unreliable or cannot be supervised adequately
B. Whenever the task involves text
C. Whenever AI is available
Reveal Answer
A. Inadequate reliability, unacceptable risks, missing permissions, or insufficient oversight may make delegation unsuitable.
Think About It: Your Delegation Blueprint
Imagine preparing a presentation for a community awareness program.
What would you do yourself, where would AI assist, and what could appropriately be automated?
Three-Part Reflection
1. Problem Awareness
What is the purpose of your presentation? Who is the audience? What activities are involved?
2. Platform Awareness
Which available AI capabilities could support research, content preparation, visual design, or document production?
3. Task Delegation
Identify at least one human-led, AI-assisted, and potentially automated task. Explain why each choice is suitable.
Prepare a short AI Delegation Blueprint showing the intended goal, the available AI capabilities, your division of activities, and the safeguards needed.
What Have You Learned?
After completing Lesson 038, you should be able to:
- Define AI Delegation.
- Explain why delegation begins with human purpose.
- Describe Problem Awareness.
- Describe Platform Awareness.
- Explain Task Delegation.
- Distinguish human-led, AI-assisted, and AI-executed activities.
- Recognize situations where delegation may be unsuitable.
- Apply the Delegation framework conceptually to an everyday activity.
PHASE 2 — APPRECIATION
Module 10 — The 4Ds of AI Fluency
Module 10 introduces four capabilities that support effective and responsible human–AI collaboration.
✓ Lesson 037 — The 4Ds of AI Fluency
Delegation, Description, Discernment, and Diligence.
● Lesson 038 — Current
What Is AI Delegation, and How Do We Decide Who Should Do What?
Lesson 039 — Coming Next
Description: How Should I Communicate?
Lesson 040
Discernment and Diligence: Evaluation and Responsibility.
COMING NEXT · LESSON 039
Description — How Should I Communicate?
Once we decide what AI should contribute, the next challenge is communicating effectively.
Lesson 039 will explore how to describe the desired product, the process to follow, and the expected performance.
How can clearer communication help AI produce more useful, relevant, and appropriate results?
Continue Your Learning
Previous Lesson — Lesson 037
What Are the 4Ds of AI Fluency, and Why Do They Matter?
Previous Module — Module 9
Automation, Augmentation, and Agency.
Phase 1 — Awareness
New to AI Fluency? Begin with six foundational modules and 24 introductory lessons.
AI Fluency Master Roadmap
Further Learning Resources
Stanford Institute for Human-Centered Artificial Intelligence
Delegate Thoughtfully. Preserve Human Responsibility.
Understand the problem.
Recognize AI capabilities and limitations.
Decide which activities belong to humans, AI assistance, or authorized automated execution.
AI Fluency begins with knowing who should do what.
