Multiple Assessment Checkpoints Responding to AI: A USask Example
In an online first-year English course, Dr. Carleigh Brady redesigned assessment to create multiple checkpoints that helped students engage with course readings, demonstrate their thinking—including orally—and receive feedback throughout the writing process. The redesign made learning more visible while adapting assessment to the realities of generative AI.
By Gwenna Moss Centre for Teaching and Learning
As generative AI becomes more common in higher education, many instructors are rethinking how they assess student learning. For Dr. Carleigh Brady, Sessional Lecturer in English at St. Thomas More College and Lecturer at the Edwards School of Business, the challenge was not simply preventing inappropriate AI use. It was helping students engage meaningfully with course readings and demonstrate their own thinking throughout the writing process.
"I start from the position that GenAI is a tool, and like any tool, can help or hinder depending on how it is used," says Dr. Brady.
That perspective shaped a redesign of ENG 112: Reading Drama, an asynchronous online course she has taught in various formats since 2016. While delivery methods have changed over the years, the learning outcomes have remained consistent: students should be able to perform a close reading of a play and use that information to formulate critical arguments in a well-written and persuasive essay.
Rethinking Assessment in ENG 112
ENG 112 is a smaller course of approximately 25 students, primarily continuing students enrolled in Arts and Science. Traditionally, assessment consisted of discussion board posts, peer reviews of essay drafts, and analytical essays.
However, COVID-19 and the emergence of generative AI prompted Carleigh to reconsider how students were interacting with course materials and demonstrating their learning. An earlier redesign in 2023 prohibited AI use, required students to document their writing process, and placed greater emphasis on originality, critical thinking, and close reading.
Initially, those changes appeared to be working. But after a difficult Fall 2025 term, Carleigh took a closer look at student performance. A review of student work suggested that some students were relying on AI-generated summaries, fabricated sources, and AI-written essays rather than engaging directly with the assigned plays. The experience led her to re-examine how the course could better encourage students to interact with readings and develop their own ideas.
With limited time between semesters, Carleigh focused on two goals: getting students to locate and engage with course readings, and incentivizing them to do their own writing. The resulting redesign centred on helping students show how they arrived at their work and providing multiple opportunities for feedback and instructor interaction throughout the term.
Helping Students Show Their Work
One of the most significant changes involved requiring students to provide evidence of where they found the quotations used in their essays. Rather than simply citing textual evidence, students were asked to submit photographs or screenshots showing the original source of the quotations they used. Those quotations had to come from authorized course materials rather than websites, excerpts, or AI-generated content.
Students received a small portion of their grade for completing the task. While the activity could not guarantee that students were reading closely, it encouraged them to engage directly with the texts themselves and made it more difficult to rely exclusively on summaries or AI-generated content.
"While I can't make a student carefully read and reflect on an assigned play, it turns out that I can make them obtain and skim through the assigned text—at bare minimum," she says. "Obviously, the hope is that, having obtained the course reading, the student will be then inspired to read it (and perhaps enjoy it!), but the bare minimum is still an improvement over not reading the play at all."
The requirement shifted attention away from the final product alone and toward the process of gathering and selecting evidence. In doing so, it created a simple but effective checkpoint that encouraged students to work directly with the assigned readings.
Building in Conversation and Feedback
A second key change was the introduction of an outline-and-chat activity designed to give students feedback before they submitted a major essay. Students first developed an outline and identified the textual evidence they planned to use. They then met with Carleigh for a brief conversation on Zoom to discuss their ideas, organization, and choice of evidence.
The conversation typically lasted 10 minutes and provided students with an opportunity to test and refine their thinking before beginning the essay. It also gave Carleigh a better understanding of how students were developing their arguments and engaging with the course readings.
The activity formed part of a larger assessment structure in which students gradually built toward more complex work over the term. Students gathered quotations, developed an outline, discussed their ideas, completed written assignments, and ultimately produced a major essay and final exam.
"The other assessments all build upon each other and allow me to check in on the students' progress at multiple stages," says Carleigh.
Rather than relying heavily on a single high-stakes assignment, the redesigned structure created multiple opportunities for students to demonstrate their learning and receive feedback along the way. For Carleigh, these checkpoints proved particularly valuable because they provided insight into students' thinking before they reached the final stages of a major writing project.
Results and Future Directions
The results were encouraging. Following the redesign, the course failure rate dropped from 21 percent to 5 percent. Pass rates also increased on both the major essay and final exam.
"I believe this improvement is largely due to the changes previously discussed—asking students to show their work, increasing supervised assessments, and incorporating them earlier in the course," says Carleigh.
At the same time, she does not view the redesign as a final solution. While the changes appear to have reduced students' reliance on AI, they have not eliminated it. Carleigh believes students are still using AI, particularly for major written assignments, although likely in different ways than before.
"Simply put, my course's current ban on GenAI doesn't match the reality of how students are using it," she says.
As a result, she is beginning to consider whether future versions of the course should allow limited AI use, provided students clearly disclose and document how they use it. For Carleigh, the experience reinforced the importance of continuing to adapt assessment practices as technology evolves.
"I hope this article points other educators to specific actions and assessments they can implement in their courses," she says. "But more importantly, I hope it gives them the confidence to try—and to fail, and to try again."
Apply in Your Context
If you want support exploring your own assessment redesign, contact GMCTL@usask.ca for consultation and workshop opportunities.
Title image credit:Zigor Agirrezabala Vitoria |Pixabay
This article was created with the assistance of AI tools, as described in the GMCTL AI Disclosure Statement.