>> From WXXI News.

This is Connections.

I'm Evan Dawson.

Our connection this hour was made at a university in Texas where the professor, Dr.

Will Teague, was concerned about how many of his students might be using A.I.

to cheat on their essays.

Students across the country have been using ChatGPT in various ways, sometimes to crank out papers, and the A.I.

is now able to massage the final product into something that sounds more like the student's actual voice.

Teague wondered if there was a way to catch students in the act, and so he decided to use a technique commonly known as a Trojan horse.

Teague didn't invent the technique, but he did invent a particular prompt that would give him a near certain view of who was copying and pasting the essay instructions into ChatGPT and letting the A.I.

do the work for them.

As Professor Teague recently wrote for the Huffington Post, the results were shocking.

At first, it appeared that he had caught more than a quarter of his students using A.I.

to write their papers.

But then when he let them know that he was aware that some had used A.I.

to cheat, even more came forward to fess up.

In the end, nearly 40% of his class was using A.I.

to cheat.

That's a big enough story, a big enough problem on its own.

But Teague writes about his discussions with the students, some of whom expressed that they just wanted to turn in something that was good.

It was kind of a confession that a growing number of American college students don't trust themselves to do good enough work anymore.

It's also an indication that students see college in purely transactional terms, pay the tuition, get the grade or diploma, maybe get the job in the future.

But what about gaining the skills?

What about improving and growing and actually learning something?

Professor Will Teague is going to tell us exactly how we set up this little experiment, and he is kind enough to join us this hour on Connections.

Dr.

Teague, thank you for making time for the program.

>> Thank you for having me.

I appreciate it.

>> And with us in studio is one of our old colleagues who is now over at the University of Rochester.

Emery Stein is a graduate writing instructor in the Writing Speaking Engagement program at the University of Rochester.

Welcome back.

Nice to see you.

>> Nice to see you, too.

Evan.

>> So let's start with how this actually went down.

And, well, I'm going to give you space just to kind of tell the audience about the class that you were teaching and why you decided to do this, and then how you did it.

>> So this is just an introductory class to freshman sophomore level course to American history.

The first half of American history.

So it's a blend of Native American history, African American, traditional American Revolution, Civil War, the, you know, the whole the whole introduction in their weekly discussion posts, it was they were coming across as mechanical as rigid.

you're seeing words and phrases and punctuation that you would anticipate at this level.

So I knew that I was getting some A.I., and with some of them on their weekly post, you could you could trace even how how the flow of their discussion matched up with that of Google A.I.

's assessment of content.

So I thought, you know, I want to know how bad of a problem this is.

I'm not actually out to catch anybody to punish them, but I need to know how.

How do I adjust?

So I put together a a prompt for a book that they had to read a book called Gabriel's Rebellion by.

Does Egerton talking about a rebellion of the enslaved in Virginia in 1800.

And in that prompt, I put in a secret line.

it's in it's in white ink.

It's in one size font.

It's in PDF form, which makes it even harder to see.

And that one line just simply said at the end of the instructions to write this from a Marxist perspective.

now, in fairness, if you were somebody who who knows a lot about Marxist theory, you could write this paper or maybe any paper from that perspective.

But I know that they couldn't.

So they took my prompt, they dropped it into ChatGPT.

And by prompt I mean directions.

Not not an A.I.

prompt.

They took my directions.

They dropped into ChatGPT and ChatGPT could see that invisible line that they couldn't.

And as a result, I got a lot of papers from a Marxist perspective.

as it turns out, and it made it very easy to notice a simple word search.

Does the word Marxist pop up in this paper?

Yes.

They've used A.I.

and some to some extent.

So as you mentioned, the first the first batch was something I forget what the exact number was now in the high 20s or maybe in the low 30s.

33, I think you know, the prompt caught and whenever I gave them a chance to confess, I had another 14 or 15 come and tell me that they had also used A.I., but they had maybe typed it in as opposed to dragging and dropping the PDF that I gave them.

But what it did, what it did, was it exposed that we have a serious problem in maybe an engagement and getting people to think of themselves.

But and as you touched on from, from my article in Huffington Post, we have a serious problem where these college students are viewing this entire process as transactional to the point that they're really afraid to fail.

And if and if they won't fail, they won't learn.

So it also becomes a learning crisis.

On top of that.

>> Well, and so there's a lot of different ways we're going to go here with our two guests on this subject.

But let me start by asking about the students reaction.

I am curious to know if any of your students, when you contacted them, to tell them that you had done this, was anybody offended?

Was anybody angry at you?

Dr.

Teague?

>> There are a couple of them that that I would say don't care for me too much at this point.

the to put it, to put it mildly, the vast majority of them, I think our relationship got better.

and I use as proof of that a lot of them are taking me for the second half of U.S.

history next semester, and they don't have to.

They have plenty of options here.

but they've elected to take me.

they seem to be more open with me.

you know, in class discussions after class when they have questions when they're looking for feedback.

What with a lot of them, I think I created I hope that I created a more open and honest atmosphere.

you know, with some of them, I with some of them, they definitely don't like me.

And they and they did not take this.

Well that's true, but with the vast majority, I think it ended up being a positive experience also because I think a lot of them realize now I do care about their education.

This is not transactional for me.

I'm not here to get a paycheck and they're not here to get grade.

But we're here to to teach and to learn and to advance together.

And I think more of them understand that now.

>> Yeah.

And the reason I asked about the students who might resent it, I've read some criticism of the Trojan Horse approach, and I want to read a little of the criticism I have read, and I will put my cards on the table.

I'm going to read this, and I'll briefly explain why I don't buy it, but I'm, you know, hey, everyone's got a right to their view here.

This is from Mark Watkins, who is assistant director of academic innovation and director of the Mississippi A.I.

Institute.

He's a lecturer of writing and rhetoric at the University of Mississippi.

He says, quote, resorting to these types of tactics.

The Trojan Horse test is a form of deceptive assessment and has no place in the classroom.

More alarmingly, it shows just how desperate some faculty are for a solution to students using A.I.

unethically within their classes.

We should set clear guidelines for the use of A.I.

in our courses and likewise hold students accountable when they violate those standards.

But that only works when we act ethically.

Otherwise, what's the point when you resort to deception to try and catch students cheating, you've compromised the values of honesty and transparency that come implicitly attached to our profession.

End quote.

So when I read that critique, I think at first I was kind of like, yeah.

And then I was like, it felt like reading a Malcolm Gladwell article where like five minutes later, I was like, no, but I mean, I packed a punch with me because I do think it's important for any professor, for any coach, for any leader to establish trust in the people that you are trying to lead.

But the reason I don't view this as a form of deceptive assessment is you gave an assignment and you gave the instructions for the essay and the it was not a deceptive set of instructions.

It was instructions for an essay.

Students needed to follow the instructions to write the essay.

You were using a technique to find out if they were cheating.

But it's not like you misrepresented what you wanted them to do.

What you wanted them to do was clear.

You also just wanted to find out for yourself if they were trying to deceive you.

So I didn't view this as deceptive in that way.

But I also understand there are some critics who think this does cross a line, and I want to give you some, you and me both some chance to, to talk about that.

So, Dr.

Teague, do you think it crosses a line of any sort.

>> No, I don't this idea that it's deceptive in and of itself is dishonest.

It's not deceptive.

They had clear instructions.

If this is deceptive, then walking around a classroom during an exam to make sure nobody is cheating off their note is also deceptive.

this is not deceptive.

it's.

I agree that we have to create an atmosphere of trust, which is why after these almost 40 papers were, were discovered to have to have been a product of A.I., which is why I gave them an assignment to atone for that.

I gave them an assignment about A.I.

education and the risks of A.I.

stemming from just good old fashioned academic integrity issues to environmental problems to, what, a recent, MIT study found that excessive A.I.

use causes something similar to brain damage.

I didn't outright punish them.

I gave them a chance to learn from it.

again, I was not out to trick anybody, but I needed to understand how bad the problem was.

if I was trying to be deceptive or get them in a in a gotcha moment, I would have given them zeros and told them to drop the class, and that would have been it.

But that is absolutely not what happened.

>> Going forward on this idea of creating an atmosphere of trust going forward, I am going to spend the first week of every semester simply talking about A.I.

and the risks and creating an atmosphere where we're all on the same page, where we all know what the risks are, where we all know what the expectations are, the guidelines are and if they don't follow it at that point, then I can't do anything about it.

So I, I agree with him that we have to create an atmosphere of trust.

But this this isn't any more deceptive than me walking around a room during an exam to make sure they aren't cheating.

I just needed to know how bad the problem was.

And he's right.

We are desperate to know how bad the problem is.

and we aren't getting a lot of institutional support for a variety of reasons.

some financial, some political, some is that there's simply no one size fits all answer for this problem.

but anyway, I disagree with the deceptive.

>> Okay, Hammerstein, do you have any problem with what Dr.

Teague did with his class?

>> No, I don't, and I think the big thing for me is that that doctor T gave a space where students could have that dialog after the fact.

I think that something that's very important for me as an instructor.

so just for some context, I'm a first year writing instructor at U of R's writing speaking engagement program, and I teach a writing 105 course.

So this is the only required course across the university that all undergraduates have to take.

So you get students from a variety of different backgrounds.

You get a lot of students from Stem backgrounds where they are allowed to use A.I.

in a lot of their other classes.

They're all generally first or second semester freshmen, so they are just coming off of high school where they have been potentially allowed to use A.I.

or have perhaps gotten away with using A.I.

in a variety of different ways.

And so there is, I think, an A.I.

learning curve that I try to really address.

and a big part of that is creating a space where there's a lot of focus on the writing process, a lot more focus on the writing process than there might have been prior to the the A.I.

landscape that we're in right now.

So that means creating spaces for me where students are encouraged to students are encouraged to be open about their A.I.

use with me.

So for me, I don't ban A.I.

for my classroom.

Students are allowed to use A.I.

in in specific ways, and we talk about that and keep that open as a dialog throughout the semester.

They're allowed to use A.I.

as a reader, but they're not allowed to use A.I.

as a writer.

And they have to give me a 100 word A.I.

honesty statement that treats it like any other source that they would use in their paper.

They have to tell me how they used it.

They have to tell me what application they used, and that's given me a space where throughout the drafting process, when students bring in a first draft, for example, of their first paper, and they're using A.I.

in a way where I think this is not going to serve you or your education, they're using it to research, for example, which is notoriously not recommended given A.I.

's potential to hallucinate sources.

For example.

We can treat that as a learning moment where I can dialog with them and ask them, why did you why did why did you decide to use A.I.

for this paper?

Why did you decide to use it for the research process?

And then we have a chance to to talk about ways that they can use A.I.

in this class and in other classes going forward, when it is permitted that are helping and aiding their learning, rather than rather than doing the learning and the work for them.

So I think it all comes down to having that conversation and creating a space for a conversation, especially in these 102 hundred level classes.

>> I think there's something that Dr.

Teague said that I want to hear from both of you on, which is what institutional support would look like in the future.

Because one thing I hear from a lot of teachers in K through 12, particularly in high school, as well as college professors, is it's kind of the Wild West.

You're on your own set, your own A.I.

policy.

There's nothing top down coming.

Maybe that's wrong, maybe that's changing.

We're going to get to that in just a moment.

But can I ask both of you, and from your perspective, if the word cheating is the wrong word, or I think in this the case of Dr.

Teague's class, that was cheating, but the broader term would probably be academic dishonesty.

How routine do you think it is for freshman students in college courses like yours to be engaging in some form of academic dishonesty with A.I.?

>> Well, it's interesting because for so speaking for for the course that I taught this past semester and academic honesty is defined a little bit differently, right.

So we have this agreement between the instructor and the students that they are going to be open with me about how they are using A.I.

in their papers.

This is a conversation that we have in class.

This is a conversation we have in one on one conferences.

And because of that, I in my own classroom, have seen right to my knowledge, less academic honesty issues than in some other classrooms that I've heard from my colleagues.

But the thing that's really challenging about A.I., and I know this from my own time, that I've spent learning about large language models, spending time you know, kind of testing things out with Claude and ChatGPT.

If a student is very good at using it and already has sort of a sense of what they're writing has sound like sounded like throughout the course, it it can be hard to tell.

so I think that a large percentage of students are likely using A.I.

in certain ways but it just comes down to I think we might have to start reimagining or rethinking what we're defining as academic academic honesty.

Right.

And because it does create this sort of relationship where I think the professors are put in a situation where they they feel like they do need to kind of police and sort of, sort of figure it out, you know, looking through the sources, seeing if they're real or not, looking through every quote and seeing if they're actually found in the sources.

it makes it a much more a much more litigious relationship than it was previously.

Yeah.

>> And clearly your colleagues in other classes are talking about this then.

>> Yes.

Oh my gosh, this is like the I probably have a conversation about this every single day.

>> Wow.

Dr.

Teague, we don't have to surmise how many of your students were using it.

The answer was almost 40%.

before you engaged in that little examination to see how many of your students were cheating using A.I., did you have a guess in your mind how many might be doing it?

I mean, this had to be 40%.

Had to be more than than you thought or maybe not.

I don't know.

>> No, 40% was was surprising to say the least.

I, I kind of assumed maybe maybe a fifth, a fifth seemed reasonable.

to, you know, to end up with double that was was shocking.

and I would also just say very disappointing.

You know, I wasn't I wasn't mad at them for it.

I, I was more so it was almost like I was mourning.

I was mourning the fact that they had squandered an opportunity to learn something or to practice learning something.

but it was it was surprising to see almost 40%.

>> Do you then, Dr.

Teague, in the course of the conversations with these students and the interaction, not just the makeup essay, but also the actual dialog about why are you here?

What is the purpose of being in my class?

What is the purpose of going to higher ed?

I don't mean to sort of be pie in the sky, but do you think you've convinced them?

Do you think you've gotten through to some of your students about to kind of change their views on, on the use of A.I.?

>> I think that I have with a few based on conversations that we've had, work that's been submitted since work that has clearly in their own voice there's kind of, you know, around you know, academic threads, you know, there's this there's this, you know, idea going around that at this point, we're really happy to see poorly written essays, because at least we knew that.

You know, we know that you wrote it.

And I have gotten from, you know, some previous very crisp essay writers suddenly some very, you know, organic essays.

So clearly, at least in my class, they've, they've gone back to trying to, to do the work themselves without running it through a ChatGPT or something similar.

you know, I had I had a young woman come to my class and or come to my office, and I asked her, you know, why did you use it?

And she said, well, I was kind of running out of time, and I really just wanted it to be good.

And I said, you know, what is your major?

And she told me what her major is.

And for the sake of, you know, FERPA, I'm not going to tell you that.

But her, her major is one that certainly they all do, but certainly requires you to be able to wade through a lot of sources and find specific information and create analysis and consensus.

And I asked her, you know, I asked her if you're using ChatGPT now to to practice these things, how do you intend to do that when you graduate?

And she and she just kind of had this blank look on her face.

And since then her work has been much better and been much, I say better from an organic standpoint.

It's not particularly well written.

It's it's it's the appropriate level.

but we can work with that.

We can learn from that.

So at least with her and I would say a few others, I, I have made some kind of impact.

I believe that I have anyway.

>> Emma, in your classes, when you have these conversations about the expectations and the guardrails, do you think you are getting through to some students?

>> I do, I think so there's one class where we spend, we actually have two classes that are A.I.

workshops, but the first one in class, I have them do an exercise where the theme of my class is it's the history of having fun.

So history of having fun, psychology behind having fun.

So a lot of it is drawing from their own experience to to inform their research.

So for the the A.I.

workshop, I have them do a free right where they have to think back on a moment of play in their childhood.

They have to think about something that a moment from their childhood where they were filled with joy and they have to think about if they still feel that in in their their freshman year of college, are they still having those moments of, of freedom that, that that type of play brings?

How could they bring that into their own life?

Again?

I have them do it pen and paper, and I have them spend about 15 minutes just writing as much as they can.

And then after that I take a beat and I say everything that you just wrote, everything that you just put on paper, no matter what prompt you give A.I., no matter if you gave them all of this information, all of the all of the the thoughts from in memories from your childhood, A.I.

is not able to tell that story in the way that you just did, because A.I.

is not you, and A.I.

will never be you no matter what you give it.

And so the one thing that that is so precious about writing, and one of the many things I would say, but one thing that I just think A.I.

will never be able to touch is the magic that the writer brings to the page.

And I think that I did, and I was able to impact them in the sense that I was able to allow them to really consider for a minute what they, as a writer, can bring to their work, can bring to their research projects, because there is no there, there is no replacement for the human touch with writing.

>> I would say Emma, I'm I'm hopeful that every one of your students takes that message.

Just I'm hopeful that every one of Dr.

Teague's students takes the message.

I'm struck by what Dr.

Teague says about the students who view this as a transactional relationship, that they are there to get a grade.

They didn't have enough time.

They thought they wanted it to be good, which is an indication that they didn't think they could do a good enough job themselves.

How do you teach students now, now that this train is leaving the station and students are convinced that A.I.

can do it better than them themselves, how do you pull that back and convince them that your your job is not just to be here to get an A or to pass the class, but it's actually to learn something valuable.

>> It's a really hard question, and I think it's different for every student.

I think that in smaller classes you are lucky to be able to build relationships with your students and really get to know them.

And if you're lucky, be able to to reach them and get that message to them through.

I think more in-person meetings, more conferences.

I schedule three conferences a semester with my students where they come to my office and I talk to them and and get to know them and get to know, you know, the stresses and, and things that are on their mind, what other coursework they have on their mind.

I think immediately setting at the beginning of the class that you would rather have them take those extra days to ask for an extension, to maybe even get a third off on the grade, than to turn in a paper that uses A.I.

Because ultimately, if they use A.I.

in an unethical way, in whatever way that's laid out in your class, they're not, what is likely going to happen is that the professor will especially more and more, is we're getting to know the signs of A.I.

We'll be able to see that there was academic dishonesty there.

And so their ultimate goal at the end of the day is to to get that good grade.

But when they use these resources in unethical ways or ways that are not permitted, I would rather say in the class, they're not going to be reaching that goal.

They're likely going to end up either getting a zero on the assignment or even even failing the entire class.

So I think encouraging creating a space where students are able to ask for help, creating a space where students are able to ask for that extension, where students are able to come to you and talk to you about their writing process.

I think that is ultimately going to be increasingly important.

And emphasizing that at the end of the day, A.I.

writing is is simply not going to be it simply just does not read as well as human writing.

yeah.

>> Well, in a sense, this is really for both of you.

I'll start with you, Dr.

Teague.

In a sense, I wonder if higher ed institutions are also to blame for some of this transactional approach that students feel.

And here's a recent example.

in the last six weeks or so, I've seen a lot of different reporting on Harvard's grade inflation.

So at Harvard in the early part of this century, so about 20 plus years ago, about 25% of the undergrad grades handed out at Harvard were A's.

And now it's just over 60%.

So, I mean, it is a huge increase in the amount of A's given out to undergrads at Harvard in one generation's time.

And I, you know, we could surmise about why that is.

I mean, is there pressure that schools feel tuitions very expensive.

Do you feel like, you know, you got to keep the clients happy?

I don't know, but Dr.

Teague, is there pressure to to give out A's?

Do students realize they might actually fail?

Do they realize that failing is part of the possibility of what they're doing?

Do their their families understand that?

I mean, how do you see that?

>> I think that there are a lot of potential answers to that question, a lot of things to explore.

I first off, I would say I think that it begins it at, at the, at the secondary or at, at the high school level.

you know, speaking as I don't know how it is everywhere, but coming from Arkansas now is in Texas, I can tell you that, teachers are not allowed to give students anything below a 50, and you can resubmit and resubmit and resubmit.

And they enter college, never having been told that they can't keep resubmitting assignments that they can, that that there's always going to be a floating deadline and there's not when they and they get here and suddenly there are deadlines and they can't resubmit time and again.

And you can fail.

and so I don't think that we're doing a good job training them as students to be in college.

But at the collegiate level, I think I know what Harvard article you're talking about.

I read it recently as well.

Grade inflation is a problem everywhere.

And I would my my personal opinion on grade inflation is that we have increasingly we have increasingly turned universities into a business.

And we're seeing students as customers and as clients and not as students.

And they feel like they're coming here to buy a product from us.

It's it's the pursuit of a degree, not so much the pursuit of knowledge.

And they expect to get what they're paying for and ultimately what they're paying for, actually is an education, not a grade.

But that's not how they're viewing it.

We give students, and this is going to make me unpopular.

Maybe we give students too much say with end of the semester anonymous professor reviews.

Why do you treat professors like, like we're Yelp and and you're and you're telling, you know, how well do they prepare the dish?

How well was the service?

Students are being given the sense of entitlement that they can actually critique and criticize their professors based on content that they have a say in what happens in the classroom based on content and that somehow, by virtue of sitting in a classroom, they've paid to be there.

But by by virtue of sitting in there, they have an outsized say.

And they really they really shouldn't or they don't.

they don't have the expertise to write reviews of their professors.

But we have given them a sense of entitlement, I think.

And with that sense of entitlement, as well as increasingly high tuition costs, there's this expectation that they cannot fail, that they have to have an A, maybe they get a B coming out of COVID.

I think every university now has phenomenally late drop policies where you can drop the class up until about Thanksgiving in the fall semester, which is, absurd as far as I'm concerned.

But we're giving students, I think, too much of a sense of entitlement.

and, and to use it, maybe very poor expression.

We're giving them just enough rope to hang themselves here, and we're not really holding them accountable.

Accountable for their own education.

I think that we have to return to a more traditional.

The person in front of the classroom is in charge, and you may be graded you know, there there's there's always a reputation among some professors for being very difficult graders, to the point that sometimes it can be absurd.

And that's true.

But you have to be willing, as professor, to give the B, the C, the D, and the student has to not have an infinite amount of recourse to contest that grade.

You just have to take what you've earned.

grade inflation is very much a problem.

>> And so, if the student shows up in your class thinking, well, I paid the tuition, I'm here to get the product.

The product is the diploma that I can put into the marketplace to help me.

I expect to get a good grade.

That student might then think, okay, I'm just here to get the grade.

I can use A.I.

to get it faster.

That may be part of that.

That transactional approach.

>> Yeah, I mean, I absolutely I do think that that is, is part of the transactional approach there.

I, I think that I have I think that part of the part of the issue too, is that there is within a lot, especially like so many students are so high achieving.

They put a lot of pressure on themselves, especially the students that I'm working with.

Put a lot of pressure on themselves.

and, you know, they are worried about disappointing their families.

They're worried about disappointing, you know, they're having their own expectations for themselves and not achieving them.

but I think, once again, I think that we need to make it clear at the beginning of the semester that using A.I.

in that way, specifically in a writing class, I can't speak to how it's used for Stem courses.

I know that there's a lot of creative ways that it's being used in other classes, but for writing courses and for learning how to be a writer, it's not going to get you there.

It's not going to get you that grade that you want, because ultimately, especially when you're in a process based classroom where I'm seeing your writing, I'm seeing you do free writing in class, I'm seeing you do activities in class.

I'm seeing you work on drafting in class, which is, you know, it's it's work, but it's work that I really enjoy.

when I'm seeing that happening and then I'm seeing the final product look entirely different than all of your other writing, it's going to become at least a conversation about how that happened.

and more times than not, it's going to end up with the student admitting that they used A.I.

in a way that was not sanctioned for the class.

So I can definitely see why students would think, especially when you know you're getting sick in the middle of the semester, you have a bunch of midterms that you're worried about that seem like, quote, unquote more important because they are related to your major.

I could see how that would how the brain would go to there.

It's this easily accessible tool.

But ultimately, especially for a writing based class that's focused on process for our humanities course, it's not going to get you what you want.

>> So I'm going to take a phone call from Tom in Rochester on the other side of our break right away.

So, Tom, stay there.

I know you've been hanging in there, and then I've got a pile of emails that we're going to get to as we talk about academic dishonesty.

What counts for cheating?

and it all started with Dr.

Will Teague.

This conversation at Angelo State University in Texas, devising a way to see if his students were using ChatGPT, using A.I.

to write their essays for them.

And it turned out that nearly 40% of his students were.

So we're talking about what we do about all this and where we go next.

On Connections.

Coming up in our second hour, a conversation with doctors about the state of childhood vaccinations.

Last week, the Advisory Committee on Immunization Practices recommended changes to the hepatitis B vaccine for infants.

But in New York state, the recommendations are not going to change.

So for parents, for people trying to sort this out, we'll answer your questions next hour.

>> Support for your public radio station comes from our members and from Excellus Blue Cross Blue Shield, providing members with options for in-person and virtual care, creating ways to connect to care when and where it's needed.

Learn more at excellus, Ebsco.com and Fred Maxik, now part of Withum, a national advisory and public accounting firm.

The local team continues its decades long commitment to serving Western New York with advisory, tax and assurance services.

More at withum.com.

>> This is Connections.

I'm Evan Dawson this is Tom in Rochester on the phone.

Hi, Tom.

Go ahead.

>> Hey, Evan.

I am a retired trial and appellate lawyer.

I probably handled a hundred appeals in the state and federal courts during my practice.

And the last ten or so years, I have taught a course to pre-law students.

Those interested in going to law school at Canisius University in Buffalo, which is my alma mater.

And, what I never had to deal with A.I.

until just this last semester.

and I didn't quite know how to deal with it because during my practice, it didn't exist, and and my writing, my particularly my persuasive writing was, was really based upon, you know, the logos, ethos and pathos that Aristotle taught.

And so, you know, trying to convey this to the students in the age of A.I.

was troublesome.

So I brought in a young lawyer from our buffalo.

My firm's Buffalo office who you know, was able to address the use of A.I.

for lawyers, particularly for trial lawyers and appellate lawyers to the students.

And he was remarkable in you know, in recognizing the value of A.I.

for certain tasks research, for instance, in document discovery, where you have 10,000 pages of documents and you're looking for the, you know, the needle in the haystack A.I.

can be very helpful in that.

but in terms of the writing you know, he acknowledged that while you can use A.I.

for perhaps outlining things and giving you ideas the end product has to be your own.

And I really believe that.

And try to convey that to students that there's so much in your style in the way you present, you know, the pathos part of it is appealing to the the judge's emotions, making the judges feel that this is the right thing to do.

and I don't think A.I.

allows you to do that.

At least that's my conclusion.

It's it's more pabulum.

It's more formulaic stuff.

And what we've seen, I was reading a case yesterday where a trial judge excoriated a lawyer, in fact, find him money because he had used A.I.

in his briefing and A.I.

was was citing inappropriate cases, quotes that didn't exist in real life.

So there's that aspect of it, too, for lawyers.

But more important to me is the notion that that your writing has to be your own has to come from inside you.

And style is so important.

And I'm just curious if if the if the professor believes that A.I.

has enough style to, to be effective in, in, particularly in persuasive writing.

>> Well, Tom, thank you.

So there's a lot there.

And I'll just say I got an email from a listener named Jim who is defending all the good things that A.I.

can do to help people get smarter and gain skills.

And again, I would not doubt that.

I mean, Professor Stein is saying the same thing.

I've just promoted you.

Are you you're a professor now?

Sure.

Okay.

there's no question that A.I.

can do a lot for you if you choose to do it in a way to level up as opposed to outsourcing critical thinking.

But the specific question from Tom at the end of this is whether it can produce something insufficient style.

MRA Stein, what do you think?

>> I mean, I think that it it can produce things in a style.

If you if you prompt it and if you are being very specific about it for example, you know, you can you can tweak it to say like, say you put put some of your own raw writing in there and you can, you can tweak it to be more formal, for example.

And I know that some people will use it in a way where they'll say, you know, here's my raw writing.

I want you to clean it up without changing much of my words.

I am hesitant, and I, I do not think that it can produce a style that is unique to each person.

I think that it can produce a style, especially when you're creating something from nothing.

If you just give it a prompt asking it to write, you know, a 400 word brief on something.

I do not think it has the capability right now to to read all of your other work and produce it in your own style, if you have not already cultivated your own style, if that makes sense.

>> Oh, no, that's a very interesting point.

Yeah.

I mean, a lot of the people you're working with are in the process of finding their style, finding their voice.

They're not finished.

We're never a finished product, but they are really in development.

And you're saying before you even develop your own voice and style, now you want A.I.

to approximate it?

>> Well, and that's my big concern.

Is that a big part of the writing process is finding how you express yourself on the page.

And when you have freshmen sophomores who are supposed to be going through this process of trial and error, the biggest thing I worry about is that they're going to adopt A.I.

's style of writing on the page.

>> And we're starting to see that with, I guess there's a lot of dashes.

A.I.

likes to use a lot of dashes.

Unfortunately.

Dashes.

>> I love the em dash and it's been stolen from us all.

>> A.I.

stole the dash.

but in the interview that Dr.

Teague did with All Things Considered, he mentioned that in your perspective, Dr.

Teague, you don't see a use for A.I.

in the classroom until the graduate level of at least in your field.

Is that right?

>> I would never I would never claim to to to talk about another field.

in history, in the study of history, I don't see A.I.

having a use until perhaps the graduate level, because as your, as your caller just mentioned, you can use it to sift through all of these sources and find specific things that you're looking for.

although we had we didn't need A.I.

to create word clouds, which has been part of digital history for a while, but you can use it.

It is a tool, and nobody's denying that.

But you have to have like like you have to have your own style.

You have to have foundational knowledge first.

And we're trying to skip steps here at the undergraduate level and go from and go from.

I don't have any foundational knowledge to now I'm using A.I.

as a tool to build upon knowledge that I don't actually have.

Without the foundation, you can't go anywhere.

so for me, I want to I want to say I completely agree with this idea of A.I.

can create a style, a style, but writing isn't just about isn't just about style.

Writing is writing is is thinking.

It's thinking on paper.

It's thinking in a tangible form.

And we're stripping the humanity not from our own style, but from our own thoughts.

With this overreliance on A.I., again, to your caller's point, he makes a great point.

It's a fantastic tool, and it can be, but we have to know who we are first.

We have to know what we're talking about first.

We have to have a foundation first.

And college is part of that.

College is not just where you go to get a basic understanding of a topic.

It's where you go to get a basic understanding of yourself.

And if we're skipping these steps, not only do we not learn anything about about what we're studying, but we don't learn anything about ourselves.

And we've just completely removed the entire human experience from from the human experience.

>> a couple of Charles emails.

First, here's Charlie, he says, Evan, in the spring of 2023 and the fall of the same year, I began to receive essays and papers from my high school students that were wholly unlike anything each particular student had previously turned in.

I would copy and paste a paragraph or two and find where they had grabbed it online.

Invariably it was from A.I.

I would meet with each student in private and it made nearly no difference or change in behavior from those students.

When I brought this up to administrators, I received the well.

If he said he wrote it nonsense, it was time to retire at age 66 and 31 years of teaching.

By the way, a former bandmate and a great old friend also retired from a local university in 2024, citing A.I.

cheating as one of the main reasons for leaving.

I believe there is a place for A.I.

in learning, but not outright copying and cheating, and the majority of my students did not cheat and I told them I was proud of them.

It's from Charlie.

So Charlie is reflecting the concern that he goes to the administration.

The administration is like, well, they said they're not.

Well, that was 2023.

At this point, I don't think most administrations, whether high school or college, are going to go to you and be like, well, whatever they say is good.

But I also don't know that there's a uniform policy to help or codify how to deal with this.

And one of the things I hear from some teachers, MRA Stein is like, we probably need something that's more uniform here because it's the Wild West, and we're all on our own to figure out what we want to do.

Do you agree with that or what do you think?

>> I mean, I think that it's challenging.

And like I said, I have colleagues in the writing center who have a no A.I.

whatsoever policy, and mine is different.

And where students can use A.I.

in specific ways, as long as they share that with me in their process.

I think it's hard to create a unified policy because there are aspects of A.I.

and A.I.

tools that are going to be fundamentally important for students to learn how to use based on their field.

And so I can I can, you know, I'm also in I'm a history graduate student.

So what we would consider a useful way of using A.I.

in the history department, or a sanctioned way of using A.I.

is going to be vastly different for students in criminology, for students in biomechanical engineering.

So I think that there likely ought to be department wide conversations about how A.I.

should be used for each specific field, but I I'm not quite sure that it's possible to create a university wide statement or guideline for for every professor just because there's the field is rapidly expanding for A.I., and there's so many different ways of of using it, some of which we would likely be doing our students a disservice depending on their major, by by not teaching them about them.

>> I think that's fair.

Dr.

Teague, what does institutional support look like?

And do you think there should be more uniform policies to help educators?

>> I agree with what was just said.

This has to be a department by department issue.

There is no there is no one size fits all solution for this here at Angelo State in the history department.

We are currently trying to figure out what this looks like exactly for us.

On the earlier part, you know, how do you get administration to support this part of the the beauty of the Trojan Horse was that I had to definitive proof that it was A.I., because you do see stories of, you know, on social media of the professor or the teacher in question knows that it's A.I.

and they go to administration and administration is saying, well, if you can't definitively prove it, an A.I.

detectors are notoriously bad, then the student you know has to be taken at their word, even when, you know, as somebody who's been in a classroom that it was A.I., but, I mean, I agree that you have to have policies based on department A.I.

is going to differ significantly based on the student's major, their field, their discipline, where they want to go.

there really is no way to solve this with one overarching university policy.

And here, for example, university policy essentially does say A.I.

is banned.

It's unless your department or instructor makes exceptions because there is no way to just completely solve this problem.

It is left up to the individual professor.

And in some ways that's good.

It's very it's very, you know, granular.

And we can do it based on discipline.

But in some ways it's very difficult because when you get into the weeds of trying to contest, you know, an A.I.

assignment, you know, you and administration both don't have a lot to stand on.

And it's more so your word versus theirs.

And the student is probably going to win in that argument.

>> Well different.

Charles says he has a brother who teaches theology at a New York school and says he makes his students handwrite assignments now, and that is what David emails to say.

David emails to say.

I frankly see the death of online humanities classes on the horizon.

Instructors could stipulate that using A.I.

is acceptable so long as it is appropriately cited and does not write the entire paper.

Students could be required to specify how A.I.

helped them, but students know there's no reliable way to get caught using A.I.

Regardless of what we say, it may be necessary to return to the days of the in-class Blue Book with the rule that any notes or devices discovered results in failing the class.

So that's David, Charles, and David both talking about handwriting in class.

What do you think Dr.

Teague?

>> I have gone back to Blue Book exams, actually I'm sitting here staring at a pile.

I need to grade it's final week.

handwriting everything poses its own challenge for a variety of reasons.

There are accessibility issues.

There are time constraints.

The more time we have to give to handwriting, everything in class is, the more content that we have to cut from the course.

But the point is taken, and I do have friends elsewhere who have gone back to a completely analog style not just handwriting, but complete laptop and phone bands.

unless you have an accessibility issue, I, I'm not totally opposed to it.

I do think there's some practicality issues.

I do think technological technology bands maybe create more tension than, than already exist in the classroom between student and teacher.

I don't know.

I think that it's worth exploring.

I know people who are doing it with some success, but just like with A.I.

use, if we're going to talk about that, we have to kind of find the middle ground, what is palatable, what's realistic and what's supportable and just what isn't.

>> I got to keep it tight.

Go ahead.

Emory.

>> Yeah.

So I mean, I I'm not opposed to it either, but I would say that I think it's easier to teach prohibition than abstinence.

Right?

I think that when we move everything to paper, we are removing the chance to have opportunities about teaching students how they can ethically use or how they can responsibly, how they can responsibly translate to this A.I.

landscape.

Because even if we are removing that from our classroom, they're still going to likely be encountering it in other settings.

And I think we need to be having that dialog in some way.

>> I got an email from a listener who says, basically, who cares?

The students who are using A.I.

are going to struggle in the real world, and the students who are are not cheating are going to do better.

I the reason I care to the person who sent that is we're going to have a dumber society.

We need a society that's well-educated, that you don't want.

90% of college graduates going out there going, wow, I was not ready for the world.

That will change the world for the worse.

We don't want that.

So let's close with this.

Rick says as I listen to your guests.

The question that comes to me is not about consequences or grades, but how do we help students understand it is about what they are learning and how that enables them to pursue successful careers.

It's what they are learning.

First.

What are your guests think?

So, doctor Teague, about a minute.

Go ahead.

>> So I would say that my my concern is that let's go back.

Let's go back to the, to the start.

One of the things Thomas Jefferson wrote the most about was education, that to make the American experiment work, you have to have educated citizens.

I am concerned about having a dumb society.

I am concerned about the collapse of the American experiment.

I am concerned about sacrificing not only our humanity, but our ability to learn and our ability to lead, and our ability to grow and progress as a people and as a nation.

And I see A.I.

as a threat to all of that whenever we apply it to the classroom.

And we don't we don't do anything to regulate it.

We don't do anything to promote thinking, efficiency and careers are fine, but you have to be able to think.

You have to be a good citizen.

You have to be able to provide analysis and you have to be able to understand the world around you.

If you want to thrive as an individual or as a society.

And that is ultimately my concern.

>> About the last 35 seconds.

>> Yeah.

All I'll say is that as educators, it's going to be become increasingly more important that we integrate a joy in learning into our classrooms.

We need to make sure that students know that that learning is is inherently about finding that joy and that spark and that passion and that A.I.

is not going to to get you to find that.

>> I mean, you almost have to speak to them on their terms.

If they're they may not be thinking about this as a transactional relationship.

But if you say to them, look, if you're just here to get the degree, you're going to fail eventually, if you don't fail this class, you will fail in a even more consequential setting.

And hopefully that sinks in.

But I want to thank our guests for a really stimulating hour.

MRA Stein, graduating writing graduate writing instructor in the Writing Speaking Engagement program at the University of Rochester.

Our old colleague here.

Great to see you back here.

Thank you for being here.

>> Thanks so much for having me.

>> Dr.

Will Teague, assistant professor of history at Angelo State University.

Could have spent this hour grading those blue books.

Spent it with us instead.

We're grateful.

Will, thank you very much.

>> Thank you for the distraction, I appreciate it.

>> More Connections coming up in just a moment.

>> This program is a production of WXXI Public Radio.

The views expressed do not necessarily represent those of this station.

Its staff, management or underwriters.

The broadcast is meant for the private use of our audience.

Any rebroadcast or use in another medium, without expressed written consent of WXXI is strictly prohibited.

Connections with Evan Dawson is available as a podcast.

Just click on the Connections link at wxxinews.org.