Univeristy of Pennsylvania
Close

Using AI to Create Assessments

09Jan’25

Generative AI

Generative AI can produce a lot of content very quickly, which can help instructors create assessments, assignments, and other teaching resources by serving as an assistant or thought partner. Using AI for these purposes is most efficient when it is effectively prompted. Visit the prompting strategies page if you are unfamiliar with this practice. 

This page discusses ways instructors might use AI tools to help create assessments. To learn more about creating assessments that ask students to engage with AI, visit our page on AI in assignment design.

Considerations for Creating Assessments

  • Harness AI’s strengths. AI’s ability to quickly generate a lot of content can enable you to modify assessments for various levels of difficulty, create practice assessments, or curate a large question bank to vary the versions of a quiz or exam that students receive.
  • Stay in the loop. AI doesn’t always execute tasks perfectly. It’s vital to review, evaluate, and edit its output to be sure that it meets your expectations and aligns with your goals for students.
  • Consider ethical implications. Before using AI to help create assessments, reflect on potential concerns for student privacy, obligations to disclose your use, and how you can model the ethical behaviors you want students to learn. Review Penn’s AI guidance and policies to be sure that your uses comply.

Exam & Quiz Question Banks

AI can be helpful for generating assessment questions, especially when building a large bank of similar items to create multiple versions. A generated bank can include exam or quiz questions, formal or informal checks for understanding, in-class polling or discussion questions, practice assessments, and more.

AI performs best when your requests are specific, so good prompting practices can help you get better results more efficiently. A common critique of AI in this context is that the questions it generates are too easy. Remember that you can iterate, or ask AI to revise its output to adjust to your needs. You should always carefully review the quality and accuracy of AI-generated content before sharing it with students.

Sample uses for AI:

  • Upload a past exam or example quiz question and ask AI to generate an alternate version that measures the same knowledge or changes the level of difficulty.
  • Ask AI to create distractors (incorrect answers) for a provided multiple choice question stem.
  • Upload lecture content or course materials that are not under copyright. Ask AI to create questions that check students’ understanding of the concepts.

Formatting questions for Canvas Quizzes

Instructors using Canvas Quizzes to deliver assessments may want to prompt AI to format questions that can be directly uploaded into Canvas. This is a multi-step process, but it may be more efficient than retyping each question.

  1. Enter the formatting prompt into your preferred AI platform, editing as needed to generate the desired assessment questions.
  2. Copy or download the result into Excel or Google Sheets, removing the header row.
  3. Save as a CSV (not CSV UTF-8).
  4. Use the QTI converter to upload your CSV, which will save as an XML file.
  5. Zip your XML file.
  6. Import the file as a QTI package to your Canvas course. Regardless of the quiz engine you use, import as a Classic Quiz. Importing as a New Quiz will not work for this file format.
  7. If you use New Quizzes, migrate the Classic Quiz to a New Quiz.

To modify questions or answer choices created by AI in this format, edits can be made to the CSV file before uploading or in the Canvas Quiz after importing the content.  As always, carefully review all AI-generated content for quality and accuracy before distributing it to students.

CETLI can help you and your team with this process. Contact us for assistance.

Assignments

AI can also help instructors create assignments and other materials to support student learning. Useful applications of AI will vary based on your course format and content. Good prompting practices can help you get higher quality output from AI more efficiently.

AI’s ability to complete repetitive tasks to produce a large amount of similar content quickly can be useful for generating different types of assignment materials. For example, you might:

  • Generate parallel versions of essay exams or project prompts to enable student choice or restrict unauthorized collaboration.
  • Create multiple, similar case studies or scenarios for students to engage with as part of an assignment.
  • Use AI as a thought partner to brainstorm ideas for revising, modifying, or creating a new assignment that meets your specific goals.

AI can also assist you in scaffolding student learning by segmenting complex ideas, making goals more transparent, or creating resources to aid students in completing tasks. Some ways you might utilize it:

  • Provide a complex set of assignment directions and ask AI to break the task into concrete steps. 
  • Upload your lecture notes or course materials that are not protected by copyright and ask AI to create outlines or study guides for students to engage with.
  • Create an overview or list of key takeaways to help students self-evaluate their understanding of the content.

These are only a few examples of ways you might engage with AI to create assignments and other learning resources. The best way to learn how AI is and is not useful is to try it. If you discover a useful application of AI in creating assessments that is not mentioned here, we encourage you to share it with us.

CETLI Can Help

CETLI staff are available to discuss ideas or concerns related to generative AI and your teaching, and we can work with your program or department to facilitate conversations about this technology. Contact CETLI to learn more.

Generative AI

More Resources