AI-Assisted Assessment Support
AI-Assisted Assessment Support
Some faculty and staff may be interested in using UIC’s secure Microsoft 365 “Copilot Chat” (which does not require purchasing an individual license through the webstore) when generating, reviewing, and revising learning outcomes, curriculum maps, or learning outcome rubrics. If interested in using Copilot chat for these purposes, please keep in mind the following when developing more successful chat prompts for learning assessment purposes:
1. Be specific and targeted in your request.
Take time to develop a thoughtful prompt that includes targeted, specific direction for the output you hope to receive. Consider quantifying and qualifying the direction you provide in the prompt (e.g., "provide no more than 5 learning outcomes that follow learning outcome design best practices"). If you ask for multiple items in a prompt, it may help to identify each task as a separate bulleted item to ensure the output you receive is complete.
2. Include specific guidance or references.
AI chat platforms rely on large language models to predict text. Therefore, including specific websites or reference documents to use as a guide for the output can be useful. For example, if you are crafting a prompt to revise your existing learning outcomes, you may want to include the current learning outcomes and also specific reputable learning assessment websites (such as APRA's webpage on developing learning outcomes) to inform the output.
3. Provide specific parameters or rules.
To increase the likelihood of receiving results that meet your expectations, consider having a section of your prompt include "dos" and "don'ts" based on what you want. For example, when writing or revising learning outcomes, it is important that they use observable and measurable verbs. Therefore, you might say to avoid verbs that are not observable or measurable such as "Students will understand..."
4. Test your prompt, reflect and revise, and then test again.
It may take some time to develop an effective AI chat prompt, especially if the first round of results are not quite what you hoped. It is helpful to reflect on the output and consider how to strengthen the prompt using the considerations above (e.g., Does the prompt need to be more specific in the format of the output? Does the prompt need to refer to more resources or websites to ensure more accurate content? Does the prompt need to include more "dos" and "don'ts" to produce useable results?).
Use of AI chat platforms should not be intended to replace knowledge and experience held by faculty and staff, including APRA staff. When using any AI platform, all AI-generated content still requires human interaction to review, revise, and finalize these materials to ensure they are relevant and appropriate. The use of AI chat prompts and their AI-generated results should not be assumed to be 100% accurate or relevant on their own.
If you are interested using AI chat prompts to support you in your learning assessment tasks, feel free to reach out to Scott Tharp, Director of APRA, for more information and resources.
This information is based on Tharp, D. S. (2026). Musings on AI’s Place at the Assessment Table. Assessment Update, 38(1), pp. 10-12. Wiley.