Critical Thinking with AI: An Interview with the Authors

Adrienne Kaufmann and Ashley Bourne-Richardson have been thinking critically about the use of AI in college courses since it started to gain traction in late 2022. They both come from a background of teaching college English. Adrienne Kaufmann currently teaches English composition at St. Joseph’s University, Brooklyn, and Ashley Bourne-Richardson taught composition for 15 years at Reynolds Community College.

It was important to them to create a brief supplemental resource that any instructor could use to teach their students how to use AI effectively and ethically.

I was particularly excited to sit down with them to talk about this resource, Critical Thinking with AI, and how it responds to the growing demand for AI guidance in the college setting.

Students Need Help with AI

Adrienne, you’ve been in the classroom before, during, and after the rollout of ChatGPT and other AI models. What has that been like as a professor?

Adrienne Kaufmann: Challenging. Confusing. Exciting. When AI first came out, I wasn’t an expert, and neither were my students. I remember that first semester after it released. I pulled up ChatGPT in class and talked about it. I put my own assignments in there for us to look at what it was doing with them. Interestingly, a lot of my students were much more unaware and uncertain with that tool than you might expect. We hear stories all the time of rampant cheating in very sophisticated ways. But I think we need to remember that there’s a barrier to entry with AI. It can be confusing. And there’s also levels at which you have to pay for this stuff. And that’s part of why I really wanted a resource like Critical Thinking with AI. I started to feel a responsibility to my students to teach them with AI and how to use it.

And did you find that when you pulled up ChatGPT in class, students were able to notice flaws?

Adrienne: Oh, yes. They were so good at it. We would all analyze the essay it gave me. Students got really interested in talking about the issues with it, pinpointing little things they saw as problems. I think students can do that in a lot of different ways. If we unlock this ability to think metacognitively, we can learn to get much more out of AI than the outputs it gives us.

Ashley Bourne-Richardson: I really want to underscore something Adrienne just said. Just because these students are young and tech savvy—or perceived as tech savvy and using AI—it doesn’t mean they have received any instruction in creating or evaluating a prompt. I think a lot of people are craving that intentional instruction. And the other thing I want to say is that I think it’s really incumbent upon us as instructors to communicate the value of what students are learning. We can show them AI is a tool that can help them elevate their learning, but shouldn’t be taking the place of their learning. To incentivize that, we really have to go back to very clearly communicating why our class is important, why the skills are important, how they’re going to be applied in our lives, and not to outsource them to AI.

Applying Critical Thinking and Ethics to AI

So you both recognized early on that there’s this growing need for instruction around AI. Why focus so much on critical thinking?

Ashley: It’s a skill that’s so important to teach here. You can’t just take whatever the AI spits out and assume it’s correct or that it meets the assignment standards. It’s one of the biggest concerns we all have with AI. That, and the fear that students will outsource any hard work to AI. The goal here is to teach them that in order to use AI effectively, they have got to incorporate their own critical thinking. They’ve got to evaluate. They’ve got to revise and make adjustments to the output. And they need to use that as a starting point for developing their own work.

‍The three chapters are broken down into understanding AI, using AI effectively, and using AI ethically. Tell me about this focus on ethics.

Adrienne: When we think about ethics in the college setting, we’re often just thinking about academic integrity. But what’s so great about this resource is that it helps students get a really broad picture of what ethical AI use is. There’s so many thorny ethical concerns with copyrighted material. And in this text, we go beyond just the academic integrity piece. Students come to the end not just knowing how to apply these tools in college, but having a firm ethical understanding of engaging with AI in the workplace and their life more generally outside of college.

AI Literacy: A Résumé Skill

Speaking of preparing students for the workplace, the webtext comes with an optional badge exam that students can use as a credential on LinkedIn or their résumés.

Adrienne: That’s right. The data is showing that employers want candidates to have AI literacy. (Check out this 2024 LinkedIn Learning report or CompTIA’s Job Seeker Trends for more on that.) We wanted to give students something they could walk away with that’s more easily shown to people than just describing what they know.

Ashley: It’s becoming a basic expectation. We know that. In a short span of time, AI is rapidly becoming a standard in everyone’s toolkit, or it’s expected to be so. Which makes it important to have a way that students can show they have proficiency in AI.

Making the Technical Understandable

Interacting with AI models feels like magic. How did you take a subject that’s so fraught with uncertainty and complicated questions and write about it in a way that all students and instructors can easily understand?

Adrienne: We used our critical thinking skills! We can’t assume anyone is starting out with prior knowledge. So we start with the foundations, instead of tasking students to evaluate outputs right out of the gate. We spend the whole first chapter showing students that AI isn’t magic. Without getting too technical, we talk about some of the big terms you hear—machine learning, large language models, tokens. By defining a few of these key terms and making sure students and instructors understand a basic explanation of the technology, you can then think a lot more critically about AI.

Introducing Students to AI

There’s so much AI content out there. Why would an instructor choose to use Critical Thinking with AI?

Adrienne: Our webtext is geared specifically toward introducing college students on how to use AI effectively and ethically. And unlike other resources, we have embedded AI models into chatboxes that students can practice with using guided assignments.

Ashley: That’s right. Guided practice really separates this resource from others. Plus, everything the Soomo webtext platform has to offer is a huge value added—built-in assignments, frequent formative assessment, Analytics (which helps you keep an eye on student progress), and of course the chatboxes Adrienne just mentioned. Those chatboxes really give students the opportunity to put into practice the concepts they’re learning about.

Practice Using AI

How do those chatboxes work? How did you manage to put several AIs directly into the webtext?

Adrienne: We’ve got a great tech team. The user experience is so seamless. Any student that accesses this webtext will be able to interact with ChatGPT or Claude directly on the page. You type into it just like you would if you were interacting with the models on their direct user interfaces. A lot of other supplemental materials will involve students having to go sign up for ChatGPT themselves. But with this, you can skip that step so students are able to immediately start chatting.

And this isn’t a knockoff of ChatGPT or Claude? This is the actual thing itself?

Adrienne: Exactly.

Can you tell me more about the assignments you designed around these chatboxes?

Adrienne: Yes, so in the first chapter, when we’re introducing students to how large language models work, a big focus of those first interactions with the chatbox are to have students see various features of it. So we have one assignment where students put the type of query into a chatbox that can often lead to a hallucination. If a hallucination appears, they get to see AI can confidently return a wrong answer. And that’s where they might begin to see how important critical thinking is when interacting with AI.

Ashley: Our other assignments teach them to be more literate with AI in that way that’s tailored to classroom scenarios. Students reflect on their interactions with the chatbox, and instructors can leave direct feedback on those reflections. They can post comments, talk with students about what they saw, and so on. We’re trying to first help students understand the features of these large language models and what they’ll see and then apply what would this look like if they were using it in their college courses.

Incorporating Critical Thinking with AI Into Your Course

How can you pick this up and integrate it into your class tomorrow? What have you done to make that integration easy for instructors?

Adrienne: For starters, the webtext integrates directly into your LMS, so it’s super easy to make it a seamless part of your course. I’m going to be using the webtext myself, and I just worked with our awesome customer support team to get it integrated directly into the course modules in my Canvas course. We also didn’t want to assume that instructors who were using this webtext were going to be experts in AI. So we designed the text in a way that could also be an entry point for instructors. We have instructor-facing pages that help outline the contents and provide guidance on creating an AI policy. And beyond that, instructors can use the webtext in class in so many different ways. We have response boards they can use to start off class discussions and practical exercises that could be the basis for small group activities. Instructors can look at the analytics and decide if they want to focus in-class time on anything students are struggling with.

Ashley: That, plus instructors can also use the webtext to walk students through their individual or institutional AI policy. At the start of Chapter 1, we talk to students about these policies and why they’re so crucial to find. Students will need to identify that policy before moving forward in the text.

That would help give students a lot of clarity. Students deal with this dilemma about whether they can use AI to help them with coursework.

Ashley: They do. And I like too that we’re really intentional about not being prescriptive. We aren’t necessarily trying to give students specific answers about what the policy should be. We give guidance about what kinds of things the policy might contain and what kinds of information they need to look for. We’re giving them a framework for questions to ask and steps to take as they move forward, which is empowering. And it reinforces our emphasis on critical thinking.

AI-Generated Images

Another thing that jumps out to me in the text is images. They’re all AI generated. And they’re completely bonkers. Do you have favorites?

Adrienne: My favorite is the smiling goldfish with the confetti.

Ashley: For me, it’s the party raccoons.

Adrienne: They’re all so fun. And we actually included all the prompts that created them in the captions. We’ve talked a lot about AI output in this conversation, but a big focus of the book is on input as well—how you’re instructing the AI. What language are you giving them? What produces what? So I think seeing those images and seeing the prompt directly below them is just one other way to get students thinking “How are we talking to the AI? And what are they generating from these inputs we’re giving them?”

‍

Developing A Webtext Anyone Can Use

Now this webtext is slightly different from other webtexts we’ve developed. For starters, it’s a mini webtext. It’s only three chapters. And it’s meant to supplement any course. Why did you decide on this format?

Adrienne: From the beginning, we wanted it to be broadly applicable. There’s no domain AI isn’t touching. And there’s no professor that doesn’t feel some obligation to teach in light of AI.

Ashley Bourne-Richardson: That pressure is felt across disciplines. We wanted any instructor in any discipline with any AI policy to be able to use this webtext. But we also needed to be really mindful that because it’s a supplement, it needs to fit easily into courses with full workloads already. We didn’t want to make it so robust that instructors would feel like it’s too much to add. So we set out to give just enough depth and enough practice to be useful. That’s how the three-chapter structure came about.

It sounds like you both have really considered the instructor as much as the student. Thank you both so much for your time today!

If you’re interested in exploring Critical Thinking for AI yourself, visit soomo.co/AI.