Co-Creating with AI: A Dance of Ideas

Assignment created by David Hicks (School of Education), using prompts with Claude AI

Other Assignments In This Series

This assignment is the second week of a four-week series focused on learning about generative AI.

Rationale

This assignment introduces students to prompt engineering and encourages them to view AI as a collaborative tool. By iteratively refining prompts, students learn to guide AI outputs effectively while maintaining their own voice and creativity.

Objectives

  • Learn basic prompt engineering techniques.
  • Use AI as a brainstorming and drafting partner.
  • Distinguish between AI-generated content and personal contributions.

Activities

  1. Read these resources to introduce key concepts and background knowledge on AI:
  2. Use these questions in small group discussions or online forums to explore the readings:
    • How does collaboration with AI change the way we define creativity?
    • What does it mean to be an “author” or “creator” when AI tools are involved in the process?
    • In what ways might AI enhance human creativity? In what ways might it limit it?
    • What ethical or professional concerns arise when AI is used in creative industries?
    • How can we tell whether a creative work is the result of human or AI effort—and does it matter?
    • Do you think using AI in the creative process undermines originality, or can it help people be more original? Why?
  3. Introduce the major project for this lesson:

    In this assignment, you’ll engage in a creative tango with AI. Choose a project that’s quintessentially “you” in your discipline. Then, use AI as your brainstorming partner, refining your prompts through several iterations. The goal is to guide the AI’s outputs while ensuring your unique voice shines through.

  4. Have students follow this schedule as they work:
    1. Project Conception (Day 1–2)
      • List 3–5 project ideas quintessential to your interests and discipline.
      • Note key components and your initial vision for each.
      • Choose the one that most excites you.
    2. AI Partner Selection (Day 3)
      • Choose the AI tool that best fits your project.
      • If unsure, experiment with 2–3 tools briefly and select the most responsive.
    3. First Dance (Day 4–5)
      • Write an initial prompt describing your project idea.

        Example: I’m creating an eco-friendly fashion line inspired by bioluminescent creatures.

      • Note AI’s response and reflect on its understanding of your project idea.
    4. Prompt Refinement I (Day 6–7)
      • Refine your prompt with more specifics and constraints.

        Example:
         Design a dress inspired by the Aequorea victoria jellyfish. Focus on translating its green fluorescent protein into sustainable, light-emitting textiles.

    5. Prompt Refinement II (Day 8–9)
      • Further refine by layering in emotional tone or deeper context.

        Example: The dress should embody ocean conservation. Include references to plastic waste. Aesthetic: haute couture meets environmental activism.

    6. Final Collaboration (Day 10–11)
      • Perfect the AI’s contribution.
      • Merge AI’s ideas with your own. Decide what to modify, amplify, or contradict.
    7. Creation & Reflection (Day 12–14)
      • Create the final project.
      • Write your 300-word reflection using dance metaphors. When did you lead? When did AI?
  5. Submit the following materials:
    • Your project
    • A log of prompts and AI responses
    • A 300-word statement detailing how you guided the AI and where you added your own touch

Model Student Responses

Prompt-Refining Examples

Environmental Engineering

Make Houston’s bayous resilient.

  • Start: Generic “flood control.”
  • Refine: Focus on local climate, soil, and cultural landscape.

Music Composition

Create a fusion jazz piece.

  • Start: Combine jazz and another genre.
  • Refine: Incorporate Indian ragas, West African rhythms, and jazz theory.

Political Science

Draft a Universal Basic Income (UBI) policy.

  • Start: General UBI concept.
  • Refine: Tailor to Sweden’s labor market and tax system.

Environmental Engineering Major’s Example

Project: Urban Green Corridor Design

AI Tool Used: ChatGPT
Prompt Evolution
  1. Initial: “Give ideas for urban green corridor.”
    Response: Generic suggestions like “plant trees” and “create bike lanes.”
  2. Refined: “Suggest innovative urban green corridor designs for a flood-prone city with hot summers.”
    Response: Better, mentioning permeable pavements and heat-resistant trees, but lacking specificity.
  3. Further Refined: “I’m designing a green corridor in Houston, TX. Summers reach 95°F, and it faces bayou flooding. Suggest bioswale designs, native tree species, and smart water management tech. Consider community needs like shade, recreation, and cultural spaces.”
    Response: Excellent. Detailed bioswale designs using native switchgrass, suggestions for live oaks and bald cypresses, smart stormwater systems with IoT sensors, and ideas for shaded amphitheaters and community gardens.
  4. Final Polish: “Great ideas! Now, factor in Houston’s diverse population. How can the green corridor reflect Hispanic, Vietnamese, and African American cultural elements? Also, suggest low-income community engagement strategies.”
    Response: Outstanding. Proposes a “cultural botany” section with plants significant to each community (e.g., Vietnamese lotus ponds), multilingual signage, and partnerships with community centers for design workshops and maintenance jobs.
Project Summary

My “Resilient Roots” green corridor in Houston synergizes AI insights with my engineering expertise and community awareness. ChatGPT’s suggestions on bioswale designs and native species grounded my technical plan. I computed the exact switchgrass density and live oak placement for optimal flood mitigation and shade.

The AI truly shone in contextualizing the corridor culturally. Its “cultural botany” idea inspired me to research further. I discovered that African baobab trees, while not native, are culturally significant and drought-resistant. I’ve included a baobab grove as a symbolic resilience monument.

However, ChatGPT missed Houston’s air quality issues. Drawing on my coursework, I added an air-quality component, strategically placing pollution-absorbing trees like American elms near highways. I also transformed the AI’s generic “community workshops” into a “Green Corridor Academy,” offering paid apprenticeships in green infrastructure to local youth.

In this “dance,” AI provided the rhythm of solid, context-aware suggestions, while I led with technical precision, cultural depth, and environmental justice steps.

Sustainable Fashion Design Major’s Example

Project: LumiNation, A bioluminescence-inspired, eco-activist haute couture line

Prompt Evolution
  1. Initial: “Create an eco-friendly fashion line inspired by bioluminescent creatures.”
    Response: Generic ideas like “glowing fabric,” “ocean color palette.”
  2. Refined: “Design a dress inspired by Aequorea victoria jellyfish. Focus on its green fluorescent protein for sustainable, light-emitting textiles.”
    Response: Better! Suggested electrospinning jellyfish proteins into nanofibers, coating recycled polyester. Missed activism angle.
  3. Further Refined: “The dress embodies ocean conservation. Reference threats like plastic waste. Use 100% recycled materials. Aesthetic: haute couture meets environmental activism.”
    Response: Strong! Proposed 3D-printed coral motifs doubling as microplastic traps. Suggested incorporating bioluminescent bacteria in biodegradable sequins.
  4. Final Polish: “Fantastic concepts. Now, make the jellyfish’s bell shape avant-garde. Add a sensor that brightens bioluminescence when near high-CO2 areas, signifying ocean acidification. How can the packaging spread awareness?”
    Response: Outstanding! Envisioned a deconstructed bell as cascading, luminous tendrils. CO2 sensor embedded in recycled ocean plastic. Packaging: a video-augmented “coral reef” box that deteriorates to show bleaching’s impact.
Reflection

In our creative waltz, ChatGPT often suggested science—like electrospun jellyfish proteins—while I led with artistic interpretation. Its coral microplastic traps inspired me, but I transformed them into intricate, 3D-lace patterns, marrying function with haute couture aesthetics. Sometimes, we were in perfect sync: its CO2-sensitive bioluminescence perfectly expressed my activist vision. Other times, I playfully dipped away from AI’s suggestions. Its standard ‘coral reef box’ became my AR-driven, decomposing art piece. Throughout, we maintained a fluid give-and-take, resulting in a line that’s scientifically groundbreaking, visually arresting, and profoundly meaningful.

Urban Planning Major’s Example

Project: Phoenix Rising: AI-Crafted Urban Renewal for Post-Industrial Detroit

AI Tool: ChatGPT
Prompt Evolution
  1. Initial: “Suggest urban renewal ideas for a post-industrial American city.”
    Response: Generic suggestions like “convert factories to lofts,” “create green spaces,” and “attract tech startups.” Ideas are textbook urban planning without context.
    My Reflection: A bland opening move. The AI offers standard tactics without grasping the city’s soul.
  2. Refined: “I’m reimagining Detroit, focusing on its auto industry heritage, Motown legacy, and 80% African American population. How can urban design honor this history while boosting economy and community?”
    Response: Much better. AI suggests:

    • Transform assembly lines into maker spaces, preserving layout.
    • Motown-inspired music education hubs in old studios.
    • Commission murals from local African American artists on industrial canvases.
    • Urban farms in empty lots, reflecting Detroit’s self-reliance.

    My Reflection: The AI starts to feel Detroit’s rhythm, seeing history not as a relic but as a living foundation.

  3. Further Refined: “Intriguing ideas. Now, let’s tackle Detroit’s ‘shrinking city’ status—138 sq mi urban area, 680,000 people (down from 1.8M). Consider:

    • 23 miles from downtown to border (transport issues)
    • 24% households without cars
    • Rising temperatures, need for shade
    • Many eager to work, lack opportunities”

    Response: Strong grasp of Detroit’s unique urban form:

    • “Mobility Hubs” in defunct car dealerships: e-bike shares, EV charging, start-up incubators.
    • “Cool Corridors”: Transform wide roads into tree-lined, shaded bike lanes, reducing urban heat.
    • “Auto-to-Agri”: Use empty city edges for large-scale urban agriculture, providing jobs and food security.
    • Propose monorail on old factory rails, an homage to production lines.

    My Reflection: Masterful. The AI doesn’t just solve problems; it weaves Detroit’s past into solutions.

  4. Final Polish: “You’ve mapped a physical renaissance. Now, infuse social rebirth. Remember:
    • Detroit’s racial wealth gap
    • Its civil rights legacy (Motown was pivotal)
    • Local wish for ‘growth without displacement’”

    Response: A standing ovation!

    • “Motown Microgrants”: Singer-named funds for Black entrepreneurs, e.g., “Aretha Franklin Soul Food Startup Grant.”
    • Civil Rights Landmark Network: Convert key sites into living museums, professionally managed by local youth.
    • “Own Your Note”: Music-themed shared equity homes. Monthly payments build ownership, like a growing music composition.
    • Data Suggestion: Track demographics of all project hires, grant recipients to ensure equity.

    My Reflection: The AI harmonizes urban planning with cultural resonance and social justice. We’re no longer just collaborators; we’re composing Detroit’s next anthem together.

300-Word Reflection

In reimagining Detroit with ChatGPT, our collaboration evolved from a halting waltz to a soulful jazz duet, where structure met improvisation. At first, the AI played standard urban planning tunes—convert factories to lofts, add green spaces. These weren’t wrong notes, but they lacked Detroit’s unique timbre.

I then shared Detroit’s sheet music: its auto industry roots, Motown’s soul, its predominantly Black population. Suddenly, ChatGPT found the rhythm. Its suggestion to transform old assembly lines into maker spaces wasn’t just adaptive reuse; it was honoring the hands that built the city. Proposing Motown studios become music education hubs showed it grasped that Detroit’s arts aren’t relics but living inspirations.

As our duet intensified, I introduced Detroit’s “shrinking city” challenge—vast spaces, sparse population, rising heat. Many would see only urban decay, but the AI heard opportunity in the rests between notes. It riffed brilliantly: “Mobility Hubs” in old dealerships, wide roads becoming shaded “Cool Corridors.” Most strikingly, it proposed monorails on factory rails—a transport solution that sings Detroit’s industrial heritage.

Our collaboration crescendoed when I pushed for social rebirth, emphasizing Detroit’s racial wealth gap and civil rights legacy. Here, ChatGPT’s response was a standing ovation. “Motown Microgrants” named for Black music icons wasn’t mere financial aid; it was cultural empowerment. Suggesting we convert rights landmarks into youth-managed museums showed deep respect for history and community. The “Own Your Note” housing idea, where payments build equity like a growing composition, wasn’t just innovative finance; it was poetry in policy form.

Throughout our jam session, the AI consistently surprised me. When I shared Detroit’s 80% Black population, it didn’t just note demographics; it saw a canvas for Black artists’ murals. Given the city’s vast spaces, it envisioned not abandonment but a stage for grand ideas like edge-city agriculture.

In urban planning, we’re taught that cities aren’t just built; they’re composed, with each element a note in a greater symphony. What struck me about our collaboration was how attuned the AI became to Detroit’s unique composition. It didn’t just solve problems; it found harmonies between past and future, struggle and strength.

This experience reshaped my view of AI in urban planning. It’s not merely a data processor or idea generator. At its best, when guided by human context and values, AI can be a co-composer—one that helps us hear the music in a city’s very streets. In Detroit’s case, that’s a phoenix’s song, rising from industrial ashes, powered by soulful, generative harmony.

Assessment Rubric

Table 1. Assessment Rubric
Criteria Excellent (90–100%) Good (80–89%) Needs Improvement (70–79%) Poor (0–69%)
Project Quality (30%) Exceptional creativity & discipline relevance Strong quality, quite relevant Adequate quality, somewhat relevant Poor quality or irrelevant
Prompt evolution and effectiveness (30%) 3+ well-crafted, increasingly effective prompts 3 prompts showing some improvement 1–2 prompts, little refinement No clear prompt iteration
Integration of AI & personal ideas (25%) Seamlessly blends AI & original ideas Clearly distinguishes both, good balance Imbalanced or unclear distinction Mainly AI-generated or no integration
Reflection statement (15%) Deeply explores AI guidance & personal touch Clearly explains both aspects Superficial explanation Little to no reflection

 

Last Updated on July 19, 2025