‘Never trust it’: Anna Mills says Be Skeptical of AI

Artificial intelligence now permeates so many aspects of our lives. From billboards to classrooms, AI is everywhere. I wanted to know how City College students should approach AI usage in their education.

To learn more, I spoke to Anna Mills, a writing instructor at City College and advocate for critical AI literacy. Mills has researched and written extensively about AI’s integration into education. She also curates AI Text Generators: Sources to Stimulate Discussions Among Teachers, a living Google Doc filled with resources for educators.

Questions and answers have been edited lightly for clarity.

What should students demand of their fellow classmates, professors, counselors (when it comes to AI-usage)?

Well, I’ve developed an OER (Open Educational Resources) textbook. It’s a free and open textbook that is sort of an orientation to AI for students, including ethical concerns, questions of: Do I want to use it at all? If I want to use it, how do I cite it? How should I use it?

And actually, I have a pretty straightforward rule-of-thumb. If your teacher hasn’t given you other directions, then I would say use it skeptically, right? Never trust it, but turn to it as you might to a tutor—who is not going to do your work for you, but they will sort of coach you through the process.

There’s (an AI) that will do the Canvas discussion post for you. For me, there’s something satisfying about typing the assignment, working hard at it, and then hitting submit.

I’m a big believer in transparency, so I think that we as teachers need to have really clear guidelines for students about what use of AI would interfere with learning, and what, if any, we woud allow. And we need to have that discussion from the beginning of the semester, and also a discussion about what happens if we think the students used AI beyond what was allowed? Because I think there is a lot of anxiety among students about being wrongly accused of AI.

Oh yeah, there’s the trackers now where (a student) actually wrote this but it’s coming up as however-much-percent AI detection.

Exactly. And sometimes it’s just the teacher who thinks and wonders if AI has been used, and wants to have a conversation with the student. So my approach is not punitive, but I do want to know what the student has learned, right?

It’s an interesting moment because so many students are anti-AI, and then so many students are using it intensively and excited about it or just feeling like they need to learn it for the workplace. (They’re) very worried about their futures, what’s going to happen with jobs and AI? So it’s a good moment to be bringing discussions into the classroom and talking across those different approaches and perspectives.

I try to make my classroom really welcoming of people who hate it and people who love it, because I have some of both.

Even within myself I have both. I think we need to learn more about it–

Yeah, it’s stunning and it’s actually everyday reality now. What we really need is regulation. We need to know when something is human and when it is AI. And we don’t have that yet, so that’s something I’ve really been pushing for … That’s a big problem for academic integrity and online learning going forward.