Using AI to Create Rubrics
Why Use a Rubric?
Rubrics are evaluation tools that set student expectations and increase grading consistency in your courses. Rubrics can be used to assess all kinds of student work, from oral presentations to creative projects. Rubrics feature evaluative criteria, definitions that distinguish student performance across a continuum, and a scoring strategy for translating these qualitative evaluations into grades.
Well-designed rubrics can help instructors assess assignments consistently and objectively, save time in grading, and provide timely and effective feedback that promotes sustainable student learning. They also clarify expectations and assignment components for both students and teaching assistants, which can reduce uncertainty and grading complaints.
For students, rubrics can clarify assignment expectations and components, help students recognize how assignments align with course objectives, increase awareness of students’ own learning process and progress, and help them to evaluate and improve their work.
How Can AI Help with Creating Rubrics?
Generative AI can help you draft a range of teaching materials, including rubrics. But remember: AI-generated rubrics should always be treated as a starting point. You’ll need to refine and customize them to fit your specific course, assignment, and learning goals.
Step 1: Define Your Needs
Before using AI, clarify what you need from the rubric. Consider the following:
Assignment Details: What is the purpose of the assignment? What skills or knowledge should students demonstrate?
Learning Outcomes: How will students show they have met the learning objectives?
Assessment Criteria: What distinguishes strong work from weaker work? What common mistakes should the rubric address?
Rubric Type: Do you want an analytic rubric (separate scores for different criteria), a holistic rubric (a single overall evaluation), or a single-point rubric (focusing on expectations and feedback without numerical scoring)?
Scoring Scale: How many levels of performance will you include? (e.g., Exemplary (3), Satisfactory (2), Unsatisfactory (1)).
Step 2: Generate an AI Prompt
Once you’ve outlined your rubric’s key elements, you can submit a prompt to a Generative AI tool. A well-structured prompt might look like this:
“Create a clear and structured rubric for an undergraduate course on [course topic]. The rubric should be formatted as a table and use student-friendly language. It should assess the following assignment: [describe the assignment]. The rubric should align with these learning objectives: [list learning objectives]. Use a scoring scale of [e.g., Exemplary (3), Satisfactory (2), Unsatisfactory (1)] and define clear descriptors for each level.”
You can also ask GenAI to help brainstorm rubric criteria. Try this:
“Based on these learning objectives, suggest six criteria to evaluate students' work.”
Step 3: Revise and Refine
AI-generated rubrics may need adjustments to fit your grading philosophy and expectations. Review the output and consider:
Clarity: Are the criteria and descriptors specific and actionable?
Fairness: Does the rubric accommodate different approaches to the assignment?
Alignment: Does it reflect your course learning objectives and expectations?
Format: If needed, ask AI to reformat the rubric as a checklist, bullet points, or another structure that better suits your needs.
Some Tips & Best Practices
Start with Clear Expectations: Before using AI, outline the key components of your rubric, including assignment details, learning outcomes, assessment criteria, rubric type, and scoring scale. The clearer your prompt, the better the AI-generated output will be.
Prioritize Student-Centered Language: Ensure the rubric is written in clear, accessible language that students can easily understand. Replace vague descriptors like "Needs Improvement" with specific, actionable feedback (e.g., "Lacks supporting evidence for claims").
Iterate and Experiment: If the first AI-generated rubric doesn’t meet your needs, refine your prompt and ask for alternative versions. You can request adjustments like:
“Make this rubric more concise.”
“Add a column for self-assessment.”
“Provide more detailed performance level descriptors.”
Test the Rubric Before Use: Before finalizing, pilot the rubric with sample student work to ensure it effectively differentiates between performance levels. Consider getting feedback from TAs or colleagues on clarity and usability.
Watch for Common Pitfalls: AI-generated rubrics may sometimes be too generic or overly rigid. Adjust them to allow for flexibility and student creativity. Ensure the rubric remains inclusive and equitable, avoiding unintended bias in language or evaluation criteria.
Use AI Responsibly and Ethically: Treat AI as a drafting assistant, not a final authority—always critically review outputs. Consider discussing your use of AI-generated rubrics with students to promote transparency in grading.