Skip to content

Assignment 3: Human–AI Collaboration

Generative AI is most powerful not when it replaces humans, but when it works in structured collaboration with them. This project explores how Human–AI collaboration can support real–world academic, professional, or everyday workflows while maintaining human judgment, verification, and responsibility.

You will:

  • Design a multi–step workflow for a Human–AI collaboration task
  • Integrate one or more AI tools into that workflow
  • Document your collaboration process
  • Evaluate the strengths, weaknesses, and risks of Human–AI collaboration

Your focus is on process, decision–making, and oversight, not just final output.

Project at a Glance

Stage 1 – Workflow Proposal (Week 9–10)

Submit: 1 Word document (.docx) on Canvas Length: Approximately 300 words

Your proposal must:

  • Describe the workflow
  • Explain the division of labor (human vs. AI)
  • Identify potential risks

Stage 2 – Workflow Design Plan (Week 11–12)

Submit: 1 ZIP file on Canvas

The ZIP file must contain exactly two files:

  1. Lastname_Firstname_Assignment3_Stage2.docx – Final Workflow Design – Design Evolution explanation (1–2 paragraphs)

  2. Lastname_Firstname_Assignment3_Stage2_Log.txt – Complete AI Design Log

Scope requirements:

  • 4–6 clearly defined workflow steps
  • At least 3 steps should be designed to involve AI assistance
  • Workflow must be realistic and implementable

Stage 3 – Implementation & Final Submission (Weeks 13–15)

Submit: 1 ZIP file on Canvas

The ZIP file must contain:

  1. Stage3_Report.docx
  2. Lastname_Firstname_AI_Log.txt
  3. Final deliverable file(s), if applicable (format depends on the workflow task)

Implementation requirements:

  • At least 3 workflow steps must meaningfully involve AI assistance
  • Each AI-assisted step must include at least 2 rounds of interaction (One prompt–response exchange does not count)
  • Workflow Documentation + Critical Reflection: 500–800 words total
  • All AI-assisted steps must be included in one complete .txt log file
  • Log must preserve full conversation history (all prompts + all AI responses)

Important

  • Do not submit separate files instead of a ZIP file
  • Do not submit multiple log files for Stage 3. You may include multiple final deliverable files in the ZIP if your workflow requires them. This does not violate the “one log file” rule.
  • Do not selectively include only successful prompts
  • File names must follow the required format

Structure

Stage 1: Workflow Identification & Proposal (Week 9–10)

Submit on Canvas (≈300 words) as a Word document (.docx)

Identify a real workflow from your academic, professional, or everyday life that you will redesign using AI collaboration.

Your proposal must include:

  1. Description of the workflow
  2. Why this task is suitable for Human–AI collaboration
  3. What the human will control
  4. What the AI will assist with
  5. Potential risks (e.g., accuracy, bias, hallucination, privacy)

Examples (not required):

  • Literature review development
  • Policy brief drafting
  • Metadata creation and verification
  • Dataset cleaning, labeling, or classification workflow
  • Code drafting or debugging with AI (e.g., Python, SQL, HTML)
  • Building a small structured dataset from unstructured text
  • Designing and refining an instructional guide or technical documentation

Stage 2: Workflow Design Plan (Week 11–12)

Submit one ZIP file on Canvas

Your goal is to design a structured and shareable collaboration workflow before implementation. You are encouraged to use AI tools to design and refine this workflow. The design process itself is part of Human–AI collaboration.

What You Must Submit

Upload one ZIP file containing:

  1. Lastname_Firstname_Assignment3_Stage2.docx
    – Final Workflow Design
    – Design Evolution explanation (1–2 paragraphs)

  2. Lastname_Firstname_Assignment3_Stage2_Log.txt
    – Complete AI Design Log

Part 1 – Final Workflow Design (Word document)

Scope Requirement

  • Your workflow must include 4–6 clearly defined steps.
  • At least 3 of these steps should be designed to involve AI assistance.
  • The workflow should be realistic and implementable.

Provide a clear outline or diagram of your Human–AI collaboration workflow.

  1. Your workflow must show the main steps of the task. For example, if you are drafting an essay, your steps might include topic selection, outline development, draft generation, source verification, and revision.

  2. You must explain where AI is used and what you will ask it to do. For example, you might ask the AI to generate an outline based on a research question or to draft a first version of a specific section.

  3. You must describe how you will check whether the AI’s output is correct or reliable. For example, if the AI provides citations, you might verify them using Google Scholar or a library database before including them in the essay.

  4. You must indicate where you expect to revise, override, or make final decisions yourself. For example, you might rewrite unclear paragraphs, remove unsupported claims, or reject fabricated sources.

Be specific. Another person should be able to understand how your workflow operates. Your workflow should be clear that you are structuring collaboration, not simply "asking AI to do a task."

Part 2 – AI–Assisted Design Log (plain text file)

Because this is a Human–AI collaboration project, you are required to document how AI contributed to designing your workflow.

Before submitting your log file, refer to Preparing Your AI Interaction Log for instructions on how to properly copy and structure your conversation history.

Submit your AI Design Log as a separate plain text file (.txt) inside the ZIP file. It must contain the complete conversation history related to designing your workflow, including all user prompts and corresponding AI responses.

The log must:

  • Clearly label every turn as User: or AI:
  • Preserve all prompts and responses verbatim (no rewriting, summarizing, or polishing)
  • Include both successful and unsuccessful attempts
  • Clearly separate different design sessions if multiple conversations were used

You may design your workflow through:

  • One extended AI conversation
  • Multiple shorter conversations
  • Separate sessions across different days
  • Multiple AI platforms

Any of these approaches are acceptable. Your submission must demonstrate clear and sustained engagement with AI by:

  • Showing iterative development
  • Including refinements and revisions
  • Preserving both successful and unsuccessful attempts

You may exclude platform–generated system messages or accidental duplicates. Minor formatting cleanup is allowed, but content must remain unchanged.

Do not submit only a polished final workflow without documenting the interaction process. Submissions that include only a final cleaned version without evidence of AI collaboration will receive reduced credit.

AI should play a meaningful role in the workflow. Minimal use of AI will not meet the expectations of this project.

AI Design Log Template (.txt): Use the following structure in your .txt file:

==================================================
Assignment 3 – Stage 2 AI Design Log
Name:
Course:
Platform used (e.g., ChatGPT, Copilot, Gemini):
==================================================

––––––––––––––––––––––––––––––––––––––––––––––––––
WORKFLOW DESIGN SESSION 1: [Title or Focus]
Date of interaction:
––––––––––––––––––––––––––––––––––––––––––––––––––

User:
[Paste your full prompt exactly as written.]

AI:
[Paste the full AI response exactly as generated.]

User:
[Follow–up prompt, if applicable.]

AI:
[Response.]

(Continue until this design session is complete.)

==================================================

––––––––––––––––––––––––––––––––––––––––––––––––––
WORKFLOW DESIGN SESSION 2: [Title or Focus]
Date of interaction:
––––––––––––––––––––––––––––––––––––––––––––––––––

User:
[Full prompt.]

AI:
[Full response.]

...

==================================================

END OF DESIGN LOG

Stage 3: Implementation & Final Submission (Weeks 13–15)

Final Submission (Week 15)

Submit one ZIP file on Canvas.

Name the file: Lastname_Firstname_Assignment3.zip

Your ZIP file must contain:

  1. Stage3_Report.docx
  2. Lastname_Firstname_AI_Log.txt
  3. Final deliverable file(s), if applicable (format depends on the workflow task)

Implementation Scope Requirements

  • At least 3 workflow steps must meaningfully involve AI assistance.
  • Each AI-assisted step must include at least 2 rounds of interaction. A single prompt–response exchange does not count as structured collaboration.
  • Your Workflow Documentation and Critical Reflection together should be 500–800 words in total. This refers to the two narrative sections titled “Workflow Documentation” and “Critical Reflection” in the Stage3_Report template.
  • All AI-assisted steps must be documented in one complete .txt log file. The log must include the full conversation history for each step, including all iterations.

In this stage, you will implement the Human–AI collaboration workflow designed in Stage 2. AI must play a meaningful role. Execute the workflow in practice and document the process using the template below. If you revise your workflow during implementation, clearly reflect those changes.

Preserve the complete conversation history for each AI-assisted step in a single plain text file (.txt). In your Stage3_Report.docx, each workflow step must clearly reference the corresponding section in your AI log file (e.g., “See WORKFLOW STEP 2 in Lastname_Firstname_AI_Log.txt”). You may use one or more AI tools (e.g., GPT, Gemini, Claude).

Final deliverable file(s) (if applicable)

If your final deliverable is fully included inside Stage3_Report.docx, you do not need to submit a separate deliverable file.

If your deliverable exists as a separate file (e.g., dataset, code, diagram, slides), include it inside the ZIP file. The file format depends on your workflow task.

Name deliverable files clearly (e.g., Lastname_Firstname_Dataset.csv, Lastname_Firstname_Code.py).

Main report (Stage3_Report.docx)

Stage3_Report.docx must follow the template below.

Human–AI Collaboration Project  
[Replace this line with a clear and descriptive title for your workflow. Delete this bracketed instruction before submission.]

Name: XXX

Before submitting your final report, delete all bracketed or instructional text.

1. Tools Used  

[List all AI tools used. If possible, include model versions. Replace this instruction with your response.]

2. Workflow Summary  

[Describe the overall task and how Human–AI collaboration was structured.  
Briefly summarize your workflow in several clear sentences.  
Delete this instructional text before submission.]

3. Step–by–Step Execution Record  

For each step, clearly indicate the division of labor and reference the corresponding workflow step in your AI log file.

Step 1: (Title of Step)

Goal of this step:  
[Describe what this step aimed to accomplish. Replace this text.]

Human role:  
[Explain what you did directly. Replace this text.]

AI role:  
[Explain what the AI assisted with. Replace this text.]

Verification or follow–up actions:  
[Explain how you checked, revised, or continued after this step. Replace this text.]

Evidence: See WORKFLOW STEP 1 in Lastname_Firstname_AI_Log.txt (Full conversation log for this step, including all user prompts and all AI responses)

Step 2: (Title of Step)

Goal of this step:  
[Replace this text.]

Human role:  
[Replace this text.]

AI role:  
[Replace this text.]

Verification or follow–up actions:  
[Replace this text.]

Evidence: See WORKFLOW STEP X in Lastname_Firstname_AI_Log.txt
(Repeat as needed. Delete this instruction before submission.)

4. Final Deliverable  

Insert your final deliverable here OR describe the deliverable file(s) included in your ZIP (file names + brief description).  
If the deliverable is not text–based (e.g., dataset, code file, structured artifact), clearly describe its structure and purpose.  
[Delete this instructional text before submission.]

5. Workflow Documentation  

In paragraph form, explain:

– How your workflow operated in practice  
– Whether your design changed during implementation  
– The division of labor between human and AI  
– Where AI added value  
– Where AI required correction or refinement  
– How verification was conducted  

Avoid very brief responses. Be substantive but concise.  
[Delete this instructional text before submission.]

6. Critical Reflection  

Write a thoughtful reflection (several paragraphs).

Discuss:

– Whether AI improved efficiency or quality  
– Where human judgment was essential  
– Whether prompt refinement meaningfully changed reliability  
– Whether you would trust this workflow in professional practice, and why  
– What governance, ethical, privacy, or reliability issues emerged  

Avoid overly short answers, but do not add unnecessary filler.  
[Delete this instructional text before submission.]

AI Interaction Log Requirements

You must submit one plain text file (.txt) containing the complete AI interaction log for your project.

Refer to Preparing Your AI Interaction Log for step-by-step guidance on preparing your .txt file correctly.

File name format: Lastname_Firstname_AI_Log.txt

The log must contain the full conversation history for each major AI–assisted workflow step, including all user prompts and all corresponding AI responses. Each workflow step must be clearly separated using labeled section headings (e.g., WORKFLOW STEP 1, WORKFLOW STEP 2, etc.) and visible separators such as lines of =====. Every turn must be clearly labeled as either User: or AI:. All content must be preserved verbatim; do not rewrite, summarize, polish, or selectively edit the interaction.

You may exclude platform–generated system messages, clearly accidental duplicate clicks, and interface elements such as timestamps, avatars, or UI formatting. Minor formatting cleanup is allowed, but the content of prompts and responses must remain unchanged. All workflow steps must be included in the same .txt file.

Example:

==================================================
Assignment 3 – Human–AI Collaboration Log
Name:
Course:
Platform used (e.g., ChatGPT, Copilot, Gemini):
==================================================

––––––––––––––––––––––––––––––––––––––––––––––––––
WORKFLOW STEP 1: [Title of Step]
Date of interaction:
Estimated time spent (minutes):
––––––––––––––––––––––––––––––––––––––––––––––––––

User:
[Paste your full prompt exactly as written.]

AI:
[Paste the full AI response exactly as generated.]

User:
[If you revised or followed up, paste next prompt.]

AI:
[Paste response.]

(Continue until this workflow step is complete.)

==================================================

––––––––––––––––––––––––––––––––––––––––––––––––––
WORKFLOW STEP 2: [Title of Step]
Date of interaction:
Estimated time spent (minutes):
––––––––––––––––––––––––––––––––––––––––––––––––––

User:
[Full prompt.]

AI:
[Full response.]

User:
[Follow–up prompt.]

AI:
[Response.]

(Continue as needed.)

==================================================

––––––––––––––––––––––––––––––––––––––––––––––––––
WORKFLOW STEP 3: [Title of Step]
Date of interaction:
Estimated time spent (minutes):
––––––––––––––––––––––––––––––––––––––––––––––––––

User:
[Full prompt.]

AI:
[Full response.]

...
==================================================

END OF LOG

Important:

  • Do not selectively include only successful prompts.
  • If you revised prompts multiple times, include all iterations.
  • Selective editing or incomplete logs will affect your grade.

Your Step–by–Step Execution section in the main report must clearly reference the corresponding workflow step in this log file.

Grading

This project is not about producing the “best” AI output. It is about demonstrating:

  • Critical AI literacy
  • Responsible collaboration with AI
  • Professional judgment of AI outputs

This project allows you to experiment with that design in a structured and reflective way.

Grading (30 pts total)

  • Stage 1 (5 pts)
  • Stage 2 (10 pts)
  • Stage 3 (15 pts)

Stage 1 – Rubric (5 pts)

Full credit (5 pts)

  • Clear and realistic workflow
  • Clear distinction between human and AI roles
  • Meaningful identification of risks
  • Submitted in correct format

Partial credit (3–4 pts)

  • Workflow present but lacks clarity or depth
  • Human and AI roles not clearly separated
  • Risks mentioned but superficial
  • Minor formatting issues

Low credit (0–2 pts)

  • Workflow vague or overly simplistic
  • No clear Human–AI distinction
  • Risks missing
  • Incorrect format or missing required components

Stage 2 – Rubric (10 pts)

Full credit (9–10 pts)

  • Workflow is logically structured and clearly sequenced
  • AI role and verification strategy clearly described
  • AI–assisted design log shows iterative engagement
  • All required components included and correctly submitted (Word document + AI design log file)
  • Correct file format and naming

Partial credit (6–8 pts)

  • Workflow mostly clear but limited detail or weak verification plan
  • AI interaction log present but minimal or not clearly iterative
  • Minor missing elements or formatting issues

Low credit (0–5 pts)

  • Workflow poorly structured or unclear
  • Little or no evidence of AI–assisted design
  • AI design log incomplete or selectively edited
  • Missing required components
  • Incorrect format or file naming

Stage 3 – Rubric (15 pts)

Full credit (13–15 pts)

  • Clear step–by–step implementation
  • Strong evidence of structured Human–AI collaboration
  • Human oversight and verification clearly documented
  • Final deliverable (regardless of format) demonstrates meaningful human judgment and clear human oversight
  • Workflow documentation and reflection are substantive
  • AI log file included and complete
  • Log file contains full conversation history for all workflow steps
  • Correct ZIP structure and file naming

Partial credit (9–12 pts)

  • Implementation present but collaboration not consistently clear
  • Reflection or documentation lacks depth
  • Minor gaps in AI log documentation (e.g., missing part of a conversation)
  • Some formatting or submission errors

Low credit (0–8 pts)

  • Limited evidence of structured collaboration
  • Final deliverable shows minimal evidence of human revision or oversight
  • Reflection superficial or missing
  • AI log file missing, incomplete, or selectively edited
  • Major formatting problems (wrong file type, missing ZIP, incorrect structure)