
Create process-based assessments
Outline
Introduction
A critical discussion is underway in higher education broadly and the arts, humanities, and writing disciplines specifically, regarding the perceived threat of generative AI to authentic student creativity.
When you think about your students using AI tools, what is your biggest fear regarding their development?
- Is it that they’ll stop learning how to struggle with an idea, or is it that the 'final product' will become a generic, polished shell with no substance inside?
To push students toward cultivating their authentic, creative voice, we can disrupt their reliance on AI outputs as the final deliverable and model how to use AI as a partner in the creative process.
We’ll demonstrate how to use Gemini as a creative partner to generate frameworks and then create rubrics that can be used for evaluation.
We will accomplish this by navigating through specific prompts to create a layered assignment.
First, you’ll instruct Gemini to generate specific project parameters to spark a creative response.
Next, we’ll task Gemini with constructing a rubric that honors the specific parameters you outlined, focusing the evaluation on how the students navigated the assignment.
Finally, we’ll show you how to give students instructions on how to use Gemini as a 'formative beta reader'— modeling how to use AI as a collaborator.
Demo
Let’s see this strategy in action.
You’re teaching a 200-level Creative Writing & Narrative Design course. Your goal is to move students away from clichés that AI and novice writers often rely on, forcing them into deeper, more original work.
We’ll use a "chaining" method to build this assignment.
Step 1: Open gemini.google.com
Step 2: Click the Canvas button
Steps 3-6: Enter the following prompts:
- Prompt 1 (Identifying the "Generic"):
I want to design an assignment about 'Betrayal.' First, generate a list of the 10 most overused clichés and metaphors people use when writing about betrayal (e.g., 'stab in the back,' 'broken heart').
- Prompt 2 (Setting the Parameters):
Now, using that list of clichés as 'forbidden territory,' write a 15-minute in-class prompt. Students must write a dialogue where one character reveals a betrayal to another. They are encouraged to use emotional and descriptive language, but if they use any of the 10 clichés on the list, they fail the exercise. They must invent a new, specific way to describe the 'friction' of this betrayal.
- Prompt 3 (Visualizing the Atmosphere):
Generate an image to inspire this scene. Show a flooded, surreal gothic library with golden afternoon light—a space that feels both beautiful and structurally compromised.
- Prompt 4 (The Assessment & Feedback Loop):
Create a 3-point analytic rubric (Originality of Metaphor, Emotional Impact, and Subtextual Tension). Finally, write a one-sentence instruction for students on how to use Gemini to 'audit' their own draft for any clichés that weren't on the original list.
Refinement
As the instructor, it is crucial that you refine the generated rubric directly in the side-by-side workspace to ensure it meets your specific standards and expectations for the assignment:
Step 7: In the Gemini Canvas workspace, use your mouse to select the text within the "Originality of Metaphor" row of the rubric table.
Step 8: In the prompt box type the following: "Add a requirement that the metaphor must feel grounded in the character's specific background."
Step 9: Review the updated output to ensure alignment with your intentions
- You can export this guide and use it directly in your class or copy the elements including the visual mood board to your presentation.
How did seeing the cliché list first change how you thought about the student's creative hurdle?
How might you introduce this process in class to model how AI can be used to support the creative process?
Practice
Now, it’s your turn. Reflect on an existing scholarly project, creative assignment, or studio challenge where students consistently struggle with genuine originality.
Your task:
Open gemini.google.com
Utilize the P.A.R.T.S. framework to architect your personal 'Constraint Engine' prompt. This process moves beyond tool usage to strategic curricular design:
- P - Persona: Act as creative director.
- A - Aim: Generate a highly unique, time-bound in-class assignment, a corresponding visual mood board, a peer-review rubric, and student-facing AI formative feedback instructions.
- R - Recipients: For undergraduate students currently sitting in a [Course Name] studio/seminar.
- T - Theme: The constraint must necessitate student reliance on their authentic lived experiences or intellectual synthesis. The rubric must exclusively assess adherence to the imposed constraint and the resultant tone.
- S - Structure: Format as a complete student-facing assignment with clear execution parameters, followed by the mood board image, the peer-review rubric, and finally, the exact prompt the student should use to get formative feedback from Gemini on their draft.
If you’re stuck, please copy and paste this pre-built example to examine the results.
Architecture & Design (The Historical Friction)
Prompt 1
Design a challenge to create a "Communal Study Space" built from repurposed 19th-century submarine parts for second-year Architecture students. Design constraints (materials, dimensions) and a rubric.
Prompt 2
Generate a visual mood board of a dimly lit, industrial study space. The architecture is composed of rusted iron, brass portholes, dials, and pipes from old submarines. The atmosphere is claustrophobic but functional, with warm, localized lighting.
Prompt 3
Students must use Gemini to "stress-test" their design by asking: "What are the 3 biggest safety or psychological flaws in this layout for a student?"
In Person Peer Review (2 mins)
Turn to a partner and share the assignment constraint, mood board, and rubric you generated.
Looking closely at the output, what worked well for your specific discipline?
What elements felt off-base or would require adjustment before giving this to students?
Virtual Reflection (2 mins)
Review the assignment constraint, mood board, and rubric you just generated.
Looking closely at the output, what worked well for your specific discipline?
What elements felt off-base or would require adjustment before giving this to students?
Reflection
We’ve explored using Gemini to build an assignment and how to model use for students.
How comfortable would you be sharing with your students that the mood board or the assignment constraints they’re looking at were AI-generated?
Beyond your own comfort level, how does that choice (to disclose or not) align with the expectations you set for their use of AI in your course?
Does being transparent about your process help model the 'AI-considerate' behavior you want from them, or does it complicate the authority of the assignment?
Gemini can expand the capacity for ideation and structured feedback, but identifying the standards and expectations for its use is critical.
An action item for you to consider is to deploy the complete assignment, mood board, and rubric generated today as a targeted 15-minute active learning module. Demonstrate for your students how AI can instigate imagination and provide feedback without usurping the role of the creator.
