Episode Transcript
[00:00:00] Speaker A: This was made by humans.
[00:00:05] Speaker B: Welcome to Austin Community Conversations, a podcast featuring discussions about the interests, backgrounds and projects animating the members of a vibrant college community.
The views, thoughts and opinions expressed are the speaker's own and do not represent the views, thoughts and opinions of Austin Community College. The material and information presented here is for general information purposes only.
The Austin Community College Community College name and all forms and abbreviations are the property of its owner and its use does not imply endorsement of or opposition to any specific organization, product or service.
[00:01:12] Speaker C: Like all higher education institutions, Austin Community College is grappling with how best to respond to the many ways that nascent generative AI technologies are changing how colleges operate and perhaps even what colleges are on October 16, 2026, ACC hosted a faculty convocation to focus on ways that AI is changing the work of teaching and learning more specific. Specifically, the event invited participants to think about the intersection of what might be called the two AIs artificial intelligence and academic integrity.
The event yielded several hours of high level discussion and insight and showcased the power of open and engaged discussion, a power that remains undiminished despite the challenges we currently face.
Today's episode was recorded during the faculty convocation and features a conversation with Brian Funken, a member of ACC's English faculty with well considered perspectives on the ways that writing instruction changes in response to technological advances.
Here we talk about the pros and cons of widely used tools and techniques for verifying the authenticity of student writing, the role that labor concerns play in thinking about AI and academic integrity, as well as strategies proposed by members of the audience. Let's dive in.
So we're going to have a conversation about the way that we think about evidence in the domain of academic integrity.
And I thought we might maybe begin by talking about some of the tools that have been name checked so far and I think that they have maybe some of their pros and cons.
So Brian, just initial thoughts, right? Your hot take if I were to say the Respondus lockdown browser
[00:03:02] Speaker D: Respondus lockdown I mean as we heard from one of our student panelists earlier today, there are potentially issues with things like you said Google lens. I believe those of you who use Respondus now there's not only screen recording microphone camera from the computer, but you can set up a second camera so you can capture a student's screen work area, hands face which because students will sometimes take a smartphone and lean it against the screen of their computer, this is actually kind of necessary and obviously it's like extremely intrusive. Panopticon George Orwell. I think no one feels very good about using it. And I was thinking about this last night also because I think about our community college students and how many of them are seeing similar types of surveillance in their work lives. If you work for Amazon or other delivery services, you're filmed the whole time you're in the car. Right. You're not allowed to do things like sing or talk or call people on the phone. And so I think both at the level of technical issues and at the level of, like, what we're showing students about what school is and our interaction with them and like, what they mean as a larger person, I have some pretty big issues. I mean, that said, I literally have a Respondus lockdown exam running right now in two online sections of dual credit British literature too. So.
[00:04:24] Speaker C: See, we are walking contradictions.
[00:04:27] Speaker D: Right.
[00:04:28] Speaker C: I might also note, right. Because I see a few librarians in the crowd.
Having spoken with many of our faculty librarians, I gather that Respondus, it poses a serious access issue to many of our students as well. Right. So many of our students aren't able to take Respondus exams unless they use ACC computer facilities which might not be open at the campus that's available to them at the time that they need it. I've learned some really funny stuff. Like in some of our accelerators, Respondus doesn't work because if people walk behind you, then the system flags that as cheap.
And so, you know, we might notice that it is a tool, but it's a tool that has some, maybe some challenges baked into it.
All right, how about AI classifiers? So detectors like Pangram and like.
[00:05:19] Speaker D: Right. So from the get go, there's been a number of observable problems with AI detectors. They tended to give higher rates of false positives to students for whom English is a second language.
And reliability. If you, if you check students work across various detectors, there's no agency regulating the claims the detectors are making about themselves. Pangram, which the University of Chicago has recently licensed, which I think is another example of just giving the money, the lunch money to the bully while they're beating you up. You know, they seem to maybe have a better product but also a better hype team than other AI detection companies. And it's a game of cat and mouse that's constantly unfolding.
[00:05:59] Speaker C: Yeah,
[00:05:59] Speaker D: let's say we're in a small window where the reliability of Pangram is high. There's already humanizers like stealth AI that are far more sophisticated than Grammarly or some of the, some of the sort of first generation humanizers. And I just don't think we're going to land in a spot where everything stabilizes and we can trust the number.
[00:06:19] Speaker C: Okay, you know, for folks like to follow this stuff. Earlier this week, a research team published a paper in Science about a new way of approaching AI detection. It's called Idea Lens. And so, so rather than the algorithm looking for patterns in the text, it configures the text into an outline and then checks that against all available outlines to see if the idea itself is likely to have been human generated or robot generated.
And I'm not going to bet too high on the success of that. But we shall see. Okay, so how about keystroke analyzer?
As the dean of the School of Humanities and Communications, I get cold call pitches from lots of companies who would like me to somehow buy their products with the money that I don't have or control.
I've noticed that all the rage has been companies trying to sell us something like this. So this is going to be some version of a tool that will evaluate the rate of keystrokes. It will monitor the version history of documents in an attempt to see whether the content, the text specifically, is entered in a way that a human might enter it. So what do you think about these?
[00:07:49] Speaker D: Actually, there was an article, an opinion piece in the Chronicle of Higher Education the day before yesterday. Two professors from Dartmouth from their government department and they said, well, if you make students run their work through an AI detector and then you make them use Google Docs or something where you can do what they call process trace analysis, and then you make them do a little follow up exam with you, it would be easier to write the paper. And of course this is totally false. If a student takes uses generative AI to create their composition, they put it through a humanizer like stealth AI. Then they take that content and they use a plugin to Google Docs like dripwriter to add it with realistic human like keystrokes, and they simply study the composition. Not only will it take less work, but you've shown that they can read and understand a text someone else has created. You don't show they've gained the skills of being able to read the text. So not only am I really skeptical about these kinds of tools, but I also think the audience for the missing faculty, I think it's administration. And what administration wants is to buy a thing and have fixed the problem. And I am incredibly opposed to that approach.
One of the things that is appealing about AI in business context is it increases efficiency and in higher education, we are deliberately trying to decrease. We're trying to slow things down because that's where desirable difficulty and productive struggle comes in.
And so I feel like admin is thinking like business leaders, not like academics. And they're like, what is the product we can buy and make everyone use? And then this problem goes away. And I think listening to someone talk earlier about pivoting to. To working with maps like. Like the ideas that will or will not work in terms of encouraging students to do. Do work, the way that they're going to gain the skills you want are deeply situated in academic disciplines. And you can't fix that by buying a program. And you can't fix that with an order from the president's office that says everyone needs to do these things.
[00:09:52] Speaker C: Yeah,
[00:09:52] Speaker D: because I am. I am an English professor. I'm not a professor of. Well, whatever.
[00:09:56] Speaker C: Yeah, I, for my own part, had a really hearty laugh the same day that I got a solicitation from Turnitin for their product. My student account got a solicitation for Dripwriter, which is a product that is essentially in an arms race with these. So I paste my essay into Dripwriter, and then it just enters it into blackboard, one letter at a time.
Very cute.
All right, so here's a claim. I'm not going to say that it's true, but it's a possible interpretation given the fact that we've got a bunch of tools and the. The reliability of these is questionable at best.
Suppose I start seriously entertaining the possibility that it's just the case that unsupervised student work just isn't a reliable indicator of mastery or competency anymore.
Is that game over for.
For instruction or. Or what? What should I do if I'm starting to seriously take that possibility?
[00:10:56] Speaker D: And it's a good question.
To go back to the piece from the Chronicle that I referenced before.
I was rooting for those professors with their method because what they were saying was that they were seeing their peers pivoting to in class writing in class exams and oral exams. And I think those are great for testing some of the skills and competencies we want students to come out with. But in a writing class, in a timed in class essay, you can test a student's ability to do timed expository writing.
But of course, writing is a. Is a form of thinking. And when it unfolds over time with multiple iterations, there are many other skills related to critical thinking. Organizing arguments, developing one's voice and style that you can't do that you can't assess right. Just through sort of that, that sort of quick fix. But I don't think we need, I don't think we should get rid of those things either. Right. I just think that one, it's deeply contextual and the sorts of things that we need to do are going to change a lot from one discipline to the next. Maybe in 10 years where we're like, okay, these artifacts don't tell us anything anymore. But a lot of. I know everyone here loves students, but the sort of collective examples across the first hour and a half of students doing this on their math and work. I will tell you, I have students in my in person classes writing in commonplace books. So basically like intellectual diaries this semester. And I pick them up every now and then when an assessment's coming in. And they're both the students are doing amazing, beautiful, time consuming hard work that they're opting to do. I open these up and students are giving me radically different kinds of responses to my reading questions and brainstorming. They're doing some of the writing in class so I can look at what they did right in front of me and at home. And it's the same voice. They're drawing cool doodles, there's rainbows and unicorns and so on and so forth. And so no, I don't think we should give up on these things. I know this is a really long answer.
[00:12:50] Speaker C: That's okay.
[00:12:51] Speaker D: But, but in my, that might mean in class writing or even in my online classes, I have students annotate stuff on paper, take pictures of it and turn it in along with their writing. That means iterative processes of, you know, for an essay, it's easy if you just have to turn in a final draft to generate it through ChatGPT or Gemini or Claude. It's hard. If you have to write a brainstorm, get feedback from me, show me you've responded to the feedback, write draft, get feedback from your peers, show me that you've responded to that feedback and then turn in a final, a final draft.
So I don't think we should give up on this stuff. But I also think that we should realize that to continue to use these assessments well, it involves a tremendous amount of labor.
[00:13:37] Speaker C: It does.
[00:13:38] Speaker D: And so putting the money towards buying the program, that's supposed to be a silver bullet. Put it towards lowering course caps for students. I mean, especially if you're teaching online, asynchronous, there's no way that you're not doing more work now than you were three years ago. It's a tremendous increase in the workload.
[00:13:57] Speaker C: Yeah, well, so maybe, you know, following up on this, you know, anytime that I've ever made a significant revision to a course, even at the assignment level, it's been a big expenditure of labor up front. Right.
But I'm finding, and I'm sure many of you are finding, that continuing to rely on the assessments that I am used to running is becoming incredibly laborious because of the amount of time that I'm spending scanning essays a third time to try and determine whether it's authentic or what have you. So, like, I'm not going to get away from some hard work to try and address this one way or the other. And you've kind of moved us toward thinking about what individual faculty might do or try as a way of moving toward authentic assessment, knowing that, you know, the tools are there, they're not hard to access, and the kind of fast tech solutions to them aren't that reliable. So one of these might be, you know, in class supervised exercises that can work in some instances. But as you said, for some things like assessing the ability to construct reasoned arguments and to, you know, write in certain kinds of ways, it might not be appropriate.
There are others that you'll see a lot of folks talking about oral exams, for example, which are interesting, but I mean, a lot of us might not have chops in that. It's something you kind of have to learn to do for the first time.
Process oriented work. Right. So ways of structuring essays, for example, where there's many times that you see the work in progress before the actual finished product is submitted. And that could be also the case for other kinds of multi stage capture capstone assessments. And in the arts or the sciences, I'm curious if anybody here is willing to maybe talk about something that you have done or tried, whether it worked or didn't, as a way of trying to modify your approach to assessment and teaching in the face of AI.
[00:16:03] Speaker E: Thank you.
I. A few semesters ago, you know, as a cumulative effect that students were just having struggle, like struggling with writing. I teach Mexican American Studies, is the course I'm going to talk about here.
And you know, I love reading essays and I love when students like have the time to write them and I have wonderful instructions on how to do that.
But they couldn't write a thesis statement, much less an intro, much less body paragraphs, much less a conclusion. And we used to be able to take that for granted. Right. Like you sign an essay and be like, here's the instructions and they turn something in. And so I reduced the response papers, that's what I call them, from three to two.
And these are all asynchronous students, so onl asynchronous.
And the first step was a thesis forum workshop in blackboard. The second step was an outline.
The first essay was decently.
Every semester is really challenging. They do not know the difference between a thesis and an introduction. But I get to intervene in a workshop. We're not in a physical classroom, but we're in blackboard. And they actually learn because these. It's our templates, right? Like we can figure this out.
And then they didn't know what an outline was and they were canceling. I require a zoom meeting with every single student with their outline. Every single one. Or I will not grade their paper. It takes a lot of time, but it's worth it.
And so they got to learn what an outline was.
And I just, I gave them something to fill in based on the instructions. Here's how the instructions work. Here's how you work on an outline. Come to the meeting even if you have zero started on your outline. I'm going to make you fill it in before I let you go further. But we're going to work together on filling in the first paragraph. You've got a thesis at the top, you've got your outline going. Then they write their paper. The second response paper goes swimmingly.
They know what a thesis statement is. They know how to take my instructions to an outline. They all come to the meeting without having to be chased.
And they all turn in papers that have started with a conversation of them being invested in each of their paragraphs and each of the things that they're talking about. It has gone from like first response papers getting C's, D's and F's to A's, B's and C's.
At the end of the semester they have a journal. It says what you know, what are the things that you're going to take with you from this class?
And most of them say working ahead on assignments.
And I take that as the biggest win of all because they don't have a concept coming to school.
Just how much difference it makes to not start a paper. How many of us get stream of consciousness papers if they're not AI stream of consciousness papers like intro one super long paragraph and a conclusion because they started 10 minutes before it started do so I. I really have loved this. Now I would say it's about 75% buy in from the students, but I will take 75%. It has really changed the course from top to bottom.
[00:19:15] Speaker C: Thanks, Jane. And I appreciate the example in the context of an ONL class because as we've seen those, those can be distinctively challenging.
[00:19:23] Speaker F: Yes. Hi.
One thing that I have, I did long before AI was go over to contract grading. Right? Because if you familiar, it's a system of saying if you do X amount of labor, I will reward that labor because I care about that, because that's what matters. And like a lot of y', all, I teach writing, right? So in the course grade structure, it emphasizes and rewards the practice. And I also offer at this point different ways for them to do practice so they can do the activities that feel meaningful to them.
So they need 1800 points to get to an A. And I have like a, like 3,500 points in the class, right. Just because it's like, oh, they could do this thing, they could do that thing.
And when you reward the practice and you show them that you care and you give them full credit even if they have not done a perfect assignment, then it really divorces the high stakes assignments and that real anxiety that students feel about grades from the actual work and it changes the relationship between me and my students in a beautiful way. Like, Eugene, I know you put a lot of work into that.
[00:20:37] Speaker C: Thanks, Gene.
[00:20:37] Speaker D: Tonio, can I go back to something that. Yeah, please, before.
Sorry.
Just a couple things struck me while you're talking. First of all, it sounds like you're doing fantastic work in that online class. But one of them is versus something like the AI detector or looking at revision histories. You're meeting with the students for like a productive reason. It's not like a hurdle they have to jump over or whatever. And you're teaching them disciplinary moves because real professional writers creating work in your discipline don't just sit down and write down what they think. They take the time to create structure and organization before they fully develop their argument.
And I think that's, that's key, like finding in whatever discipline you're in what are authentic check ins that are actually building, that are strengthening your teaching. You know, my online students, who I have them annotate by hand on paper before they take a picture and upload it with their response. It's not just so I have the response seat, although that's very helpful. It's because one of the central disciplinary moves in literary studies is annotating on the page is talking to the work that you're thinking about. So I think they're doing better work that way. And then the other thing that occurred to me, of course, is that this is a tremendous amount of work.
[00:21:48] Speaker G: I changed one of the assignments to be an oral assignment. I teach Spanish, so I named it Spanish in my community.
And it's just a three minute, three slide assignment with very, very limited instructions. And that was by design.
I'm also extremely lenient with grading that assignment. And I have found that people not only want to do the assignment, they ask me, can I do other oral presentations, because they are so excited about what they can present. And also the way that they present information is just fantastic.
I found a way to motivate, to not only write what's on the slides, not only present that, but also actually practice a language.
It was one of the biggest things that I found this past summer. I'm very excited about it.
[00:22:47] Speaker H: Hey, I just wanted to say your classes sound amazing.
I'm sure they are.
I use revision history, but not necessarily just to see keystrokes. There are some features in it that identify how many discrete writing sessions were involved and how long they were. So I build that into my rubric. I say writing if. If I really want to help break down that usual habit of waiting to the last minute and churning something out, which I don't think is the best writing practice, I. I kind of require them to show that they have worked over time. And that's one way of that they can demonstrate that by having these discrete writing sessions. And I'm upfront with them. I say, this is how I'm using it. I want at least three separate writing sessions of at least 20 minutes each. I mean, I try to get as specific as I can so that they know what the expectation is. But I also give them a space for the mess of writing. Because writing doesn't just pour out of people's heads beautifully onto the page. At least it doesn't for me.
And so I use tabs in a Google Doc, and one of the tabs is.
Is mess.
Like, it has to be super messy. I don't want it to be pristine. I just want wor and phrases. Don't bother editing it.
So I don't know whether that helps mitigate the use of AI or not, but at least it gives them the space for the mess of thinking and writing, that it doesn't all have to just be in these lovely sentences and paragraphs yet. And then I kind of model that for them. I show them, here's something I was writing and here's what it looked like when I started.
And it's super messy. You know, write two words, cross one out. So I think it's helpful to give them room for that. And that's part of their score for the. The project. I want to see the process, I want to see the mess, and I want to see that develop over time. So
[00:24:43] Speaker C: thanks, Susan. I think one more.
[00:24:47] Speaker G: Hi, I'm Jeffrey Hedrick. I'm in communications and studies. I teach public speaking.
And the revision that I made since I started here was between the informative speeches and the persuasive speeches. I made the formative speeches, group teamwork.
And I had broke them up into groups of four or five, and they had to create a group outline.
And of course, I told them that all of them would have to talk and present and they would have to reorganize and present in a certain order.
And I also, in the informative, I gave them a choice between nine topics, and there was usually six teams.
Okay. So that. That minimized the critical thinking a little bit for them and made it a little easier for them to understand how a speech is supposed to go down. And the baby step to that was in the persuasive speech, they all have to do their own preparatory outline and they have to choose a topic.
And that was the baby step that I used help assist the students. And a lot of students, even though they didn't get the importance of the informative outline, the group outline initially when they got up and presented and they weren't doing it in order or something was going on, then they started appreciating it more.
[00:26:07] Speaker C: Thanks, Jeff.
No, I don't want to. I don't want to cut the conversation off soon, but I also want to make sure that we. That we get lunch as promised. So what I'm going to invite folks to do is to use the time that we're going to have in a minute to continue this conversation. Right? So there's a temptation that I often have at gatherings like this. Once lunch hits, it's like it's bullshit time, right? This is the time where I'm going to start telling you about how hard my week was, and I'm going to tell you about why I think that X initiative is a bunch of nonsense and your life is your own. You can have that conversation if you want. But we have a unique opportunity here to sit with people we don't always get to sit with and continue talking about this. So if there are things you've tried in your class that you haven't had a chance to share via microphone yet, you can still share them, and I encourage you to do so. So I'll close this conversation with Brian out first by thanking you, but also by noticing that we focused a lot on things that faculty might do. But of course, you know, Brian, you already talked about things like course caps. Right. Which is not something that any individual one of us is in a position to change. But that could be an institutional response to this that, you know, is at least something we could look at.
Other things might, you know, include looking at the. The prevalence of ONL classes, why we continue growing those as rapidly as we do when we know that they have these kinds of challenges baked into them.
We might look at testing centers, the capacity that is available there or that's not available there.
Many of us have anecdotes about walking by empty testing centers and wondering whether or why we might be able to use those as ways of conducting assessments and things that we might have seen in terms of the proctoring that's available via distance. Proctoring already. Right.
There are ideas in this room and I, you know, introduced that participant journal at the beginning of the session. You know, if you've got ideas, please feel free to share those because I'd like to know what they are. I'm learning from you and hopefully you're learning from one another.
So with that, thank you, Brian. If you have any concluding remarks here.
[00:28:14] Speaker D: Yeah, if it's all right, I want to throw one other idea onto the list you just ran down, which is that. So ACC has a, I think, seven part framework for an AI policy that every instructor is supposed to put in their syllabus. Right. Not one of those parts addresses evidence.
Not one. Right. And I think another resource for faculty would be standards of evidence determined at the department level.
So thinking about the discipline and thinking about evidence there, because I think otherwise we can spend a lot of our time chasing after evidence that doesn't pass muster. Right. And that is not helpful. And so everyone seems to agree that our workload has gone up. And I think some disciplinary conversations about evidence can make us more productively think about how we're engaging with and supporting our students. Authentic learning.
[00:29:06] Speaker C: Thank you. Yeah. The conversations that we have with one another, I really can't overstate the importance of those because they can help to drive policy in terms of us maybe arriving around collective agreements about evidence and standards of evidence. But they're also going to do a lot to shape normal norms, and norms are often going to do quite a lot more than policy to really determine the way that instruction looks at acc. Moving forward.
