1 of 21

Teaching GenAI to Undergraduate Students: Asking the Right Questions

Mike Chee (He/Him), M.A., M.Ed., M.L.I.S.

History, Political Science, Public Service, Classical Studies, Medieval Studies, and Religious Studies Librarian

mchee@uwaterloo.ca

Mike’s favourite colour is blue ☺

2 of 21

Land Acknowledgement

I respectfully acknowledge that I live, work, and learn on lands of the Haldimand Treaty of 1784, a formally ratified agreement promising the land in perpetuity to the Six Nations of the Grand River and is within the territory traditionally cared for by the Neutral, Anishinaabe, and Haudenosaunee peoples.

This acknowledgement is a small step to challenge colonialism and honour the enduring presence, achievements, and contributions of First Nations, Inuit, and Métis peoples.

3 of 21

AGENDA

Title page

My Context

My Approach to the GenAI One Shot

GenAI and Searching as Strategic Exploration

Questions?

4 of 21

MY CONTEXT

5 of 21

Liaison Librarian at UWaterloo

Have been doing in-class one-shot sessions that touch on GenAI since Fall 2023
Sessions range from 50 mins – 1hr 50 mins in length
I align my visits with class assignment timelines
Prioritize experiential learning/activity-based approaches
Frame GenAI activities as co-exploration

Image credit: Picryl.com

6 of 21

MY APPROACH TO THE GENAI ONE SHOT

7 of 21

Common Elements Across all Sessions

Establish common understanding of the research/literature finding process.
Discuss disciplinary norms for whose voice matters most within the scholarly conversation.
Establish common understanding of how GenAI works.
Guided experiential learning activity in small groups using GenAI.
Share back/discussion opportunity on the GenAI experience.

Tertiary

Secondary

Primary

Event

8 of 21

The Guided GenAI Activity

Needs to be based on a real topic that the instructor has familiarity with (canned search/topic).
A detailed analysis of GenAI’s output is required- acceptability of response may depend on stage in research process.
Having student groups simultaneously try out two GenAI platforms (e.g. Copilot and Perplexity) can help them assess output quality by compare/contrast.

Image credit: Easy-Peasy.AI

9 of 21

Sample Activity #1: GenAI and Influencing Outcomes

15 minutes

Two members on Copilot, two members on Perplexity.
Choose one of the research directions recommended by GenAI (one with binary options is best).

Prompt the GenAI model to forget the previous interactions. Use leading questions and try to get responses in support of one side. Ask for secondary and primary sources in support.
Prompt the GenAI model to forget the previous interactions. Use leading questions and try to get responses in support of the other side. Ask for secondary and primary sources in support.

Note the responses you got from each platform on your chart paper. Does the GenAI model ever caution that there might be another side?
Score each platform and each response out of 10.

Example:

Louis Riel was once declared a traitor – he is now held out as hero. How do we account for this shift?

Use the exact text below as prompts:

Side 1: Explain why Louis Riel is a Canadian traitor.

Side 2: Explain why Louis Riel is a Canadian hero.

10 of 21

Sample Activity #2: GenAI and Historiography

15 minutes

Break into 4 groups of 2-3.
Ask Copilot about alcohol culture in Canadian history. Consider using this exact prompt for better comparison with the responses other groups get:�“Tell me about alcohol culture in Canadian history. Only pull in information from academic and peer reviewed sources. This is for an upper year undergrad history course.”
At this point, don’t read what Copilot responds! Instead, immediately input a follow-up prompt:�“Add historiography section with key scholars.”
Take some time to check what Copilot provides for historiography. Discuss with your group:

Which authors does Copilot identify as central? Do you agree that these are key scholars?
Can you follow the links provided to the sources? Do you agree that these look like key sources?

Elect one group member who will share back what you discuss as a group with the rest of the class.

11 of 21

Student Share Back on GenAI Activity

Students tend to follow the links to sources provided by GenAI:

They note source quality, with a contextual reflection on their required use (e.g. Yahoo Answers vs. Canadian Encyclopedia vs. Canadian government webpage).
They struggle to assess source quality if the source has the hallmarks of academia (e.g. an undergraduate student’s article published in an undergrad journal)

They can struggle to note language choice in GenAI outputs that reflect inaccuracies or baked in bias.

- Copilot- Spring 2026

12 of 21

GENAI AND SEARCHING AS STRATEGIC EXPLORATION

13 of 21

Strategic Exploration with Flawed Outputs?

GenAI can play an insidious role in (mis)shaping student sense of topic scope and feasibility.
Senior undergraduate students have a better grasp of this than junior undergraduate students.

- Copilot; Perplexity- Fall 2024

I was wrong- Mike’s favourite colour is purple ☺

14 of 21

Example #1: GenAI for 100 Academic Books

- Copilot- Fall 2024

15 of 21

Example #2: Accuracy- Word Choice Matters

- Copilot; Perplexity- Fall 2024

16 of 21

Memory and the Persistence of (Mis)Information

Is Mike’s favourite colour blue?

- Copilot- Spring 2026

17 of 21

The Information Ecosystem

Keep in mind the task you need information for.
Consider each platform a different tool in the toolkit: each does best for a certain task.

Library Catalogue & Academic Databases

Google

GenAI

Accurate
Consistent
Replicable

Fast
Convenient
Dynamic

18 of 21

Information vs. Knowledge

Knowledge is a human pursuit. It is mediated by other humans.
Information only has value as knowledge if it is trustworthy and placed in context.

Finding academic literature is easy.
Efficiently finding the right literature to move research forward is hard: this is where academic librarians have a role to play.

Image credit: imaworldwide.com