Strategies for integrating GenAI into assessment design

One way to teach students about ethical use of GenAI in academic work is to model and scaffold appropriate use. While principles of appropriate use may vary from course to course and instructor to instructor, it is possible to demonstrate how GenAI can be used to augment students' learning process without relying on it for cognitive offloading. Designing specific and appropriate tasks can model how students can deepen their own learning.

Example strategies

• Asking students to use GenAI to help them develop study questions or give them feedback on writing or problem sets. TAAFT has developed a prompt, available online at no cost, that can be uploaded into GenAI to create a “Socratic Questioner” whose role is “to guide users to their own clarity and insight through a structured series of targeted, open-ended questions.” 
• Encouraging students to work with GenAI as a thought partner in the beginning stages of a project to get more ideas or explore new directions. 
• Having students evaluate GenAI output against certain criteria. In these examples, students feed the assignment prompt into AI then evaluate its accuracy and comprehensiveness and identify biases. Students can also be asked to reflect on the process. 
• Incorporating a reflection component in projects and portfolios that describe processes using timelines, maps or images or written statements which can include annotated AI chats. 
• Inviting students to write a first draft in class then allowing students to use AI to get feedback to “level up” their assignments. Students must also submit a reflection about what they changed from the draft and how using AI helped their learning. 
• Organizing students into groups to collaborate on a project where after all members contribute their ideas, they ask GenAI to organize and synthesize the ideas into a plan, then review and modify the plan. 
• Requiring students to co-create with GenAI, for example in a coding assignment or data analysis project. 

External resources

For more ideas on how to design, modify or adapt your assessments:

• The University of Toronto provides example strategies for each of the distinct positions, be it prohibiting, allowing or inviting GenAI into your assessments.
• The University of Waterloo has assembled a Guide to Assessment in the Generative AI Era that lists common assessment types and suggests ways to both avoid students’ inappropriate use of GenAI and encourage them to engage responsibly with GenAI.
•  The University of Windsor offers examples of specific types of process-based assignments, the specific challenges that might be encountered in a world of AI, and strategies for creating effective prompts to further support learning.