Transcription of the episode “ChatGTP: Cheating optimizer or force for teaching transformation?”

[00:00:15] Amy H-L: I’m Amy Halpern-Laff. 

[00:00:16] Jon M: And I’m Jon Moscow. Welcome to Ethical Schools. Our guests today are my son, Lev Moscow, who teaches history and economics at The Beacon School, a public high school in Manhattan, and Richard Miller, who retired after 28 years teaching middle and high school history in New York City schools, including Beacon and Central Park East Secondary School or CPESS, and is currently tutoring. They’re speaking for themselves and not for the New York City Department of Education. Welcome, Lev and Richard. 

[00:00:32] Richard M: Thanks for having us. 

[00:00:33] Lev M: Yeah, I’m excited to be here. 

[00:00:35] Amy H-L: We want to talk to you today about chatbots, specifically chatGPT, the new artificial intelligence software that can generate essays that are frequently indistinguishable from the work produced by humans. There have been a lot of questions of what this new technology means for teaching and learning. Richard, what was your first reaction when you learned about chatGPT? 

[00:01:01] Richard M: That’s a great question. I think my first reaction was a lot of students are probably gonna use it to get over on their teachers late at night when they haven’t done their homework. And my second reaction was, it’s probably, like most things, very complicated and will raise some issues that might be helpful for teachers, pose some controversial problems for teachers and students, and finally that I’m glad I’m retired and don’t have to deal with it. I think it’s gonna be complicated. 

[00:01:38] Jon M: Lev, how, if at all, will this influence your teaching?

[00:01:45] Lev M: Yeah, it’s a big question. I have to say that we are fairly lucky at our school because we don’t have to take the Regents, in the history department at least. And that means that we have a lot of freedom, a lot of flexibility with what we teach and, more importantly, how we teach. So what we try to do is we try to design three week to one month long units and the units culminate in some kind of assessment. And sometimes the assessment is an exam that we make up. Sometimes the assessment is an exam that we make up plus some kind of paper. Or they write a play and they perform the play, or they make some kind of art project. There’s a variety of ways that the kids can show that they’ve mastered the content. And for the most part, I don’t think chatGPT interferes with what I’m doing or what I should be doing.

I think my, my first response was holy cow. And then after thinking about it a little bit, I’m actually not that worried personally, but I think that as most schools don’t have the kind of flexibility that we do at Beacon and in the Consortium in New York City, which we can talk about more if you’d like, yeah, I agree with Richard. I think a lot of kids are going to cheat, and that’s one reason why I think that that schools should be doing things differently. 

[00:03:07] Richard M: If I could just add one other thing. I spent a lot of time today reading the voluminous number of articles already that have come out about the impact on teaching, and I was struck by how quickly this new dynamic has entered into the sphere of what teachers have to grapple with. And I think that’s kind of indicative of how difficult it is to be a teacher right now. And I think that on balance, at least from what I’ve read and thought about, I think it poses incredible challenges as well as some real opportunities that we can talk about. There were a couple of real positive things that it seems to me could come out of this. You want me to expand on that? 

Well, one of the things that struck me, I was out on a two hour long bike ride today, so I listened to a couple of podcasts about this and was thinking about it, and then was discussing it with Melissa, my wife, when I got back, is the ability of a teacher to use the program to give students feedback on their writing, which is a huge problem for teachers at a school like Beacon where you might have 60 or close to 70 students writing a draft of an essay or part of an essay, and it takes a long time to get that feedback back to them. And apparently chatGPT can give them some of that feedback on let’s say their introduction or their thesis statement. You could get students, all of them, the feedback back to them the very next day. Now, it might not be as good as Lev might be able to do, but it might take Lev a week to do that. And meanwhile, students could get feedback almost instantaneously. They could even get it in class. They could write their introduction and then get feedback. I listened to a podcast where a teacher in Oregon was talking about how already she’s been using that, and it could be helpful. Obviously there are all kinds of problems. Is the feedback as personalized and as good as Lev could give in his classroom? But even a really good teacher like Lev, it’s hard to get the feedback to students real quickly. And being candid, we have lots of colleagues who take months to get essays back to students for a variety of reasons. They’re overwhelmed. They’ve got other things. If the program can give meaningful feedback, that could be helpful. 

[00:05:42] Amy H-L: Lev, you mentioned that the advent of the chatbots points out that you should be teaching differently. What did you mean by that? 

[00:05:54] Lev M: The other day we were having a department meeting and people were justifiably nervous about the future with chatGPT. And we were talking about this becoming arms race. There’s other AI which already can flag AI generated essays. And we were talking about just what our response should be and how punitive we should be when we catch kids. A couple of teachers already caught kids cheating. And then, so the discussion in the department was, well, what do we do now? And I think it’s a legitimate conversation to have.

I’ve also been thinking a lot about people who’ve been coming to the United States as asylum seekers and that they’re being bused, mostly from Texas, but now also from Colorado. I think Florida was doing it for a while, too, and many thousands of people, I think it’s over 35,000 people this year, have been bused to the Port Authority in New York City, which is just two blocks from our school, and I’ve been doing a little bit of work at the Port Authority, a little bit of volunteer work. And you know, it struck me that, again, our school’s a little bit different and the 30 or so Consortium schools in New York are a bit different, and we can talk about what the Consortium is and why I think it should expand, but it should thoughtfully expand. I don’t think that everybody should right away be doing portfolio-based assessments or performance-based assessments, but I think if more schools did, it would be a better, better world, a better education system. But one of the things that we could do, I was thinking, is that we have a con law class in our school. Con law class is taught by a really talented colleague, and 

[00:07:40] Jon M: Con law. Is that constitutional law?

[00:07:42] Lev M: Yeah, constitutional law. Sorry. And it’s taught by a really wonderful colleague and brilliant guy. And he does a great job. And I was thinking that. Why not, and I’ve been thinking a lot about Neil Postman’s book, The End of Education, and the whole last third of the book is about ways that we could transform education. And I was thinking why couldn’t we be doing lots of work with these asylum seekers two blocks away from our school? And there are all sorts of constitutional questions that come up. For example, what right does a state have to send people from one state to another state? What is the role of the federal government? When you get to the Port Authority, you see that the National Guard are there guarding the asylum seekers and the volunteers, but the Port Authority police are just outside of the barrier. And the New York City police are not there at all. I think the question of what right asylum seekers have to enter the country is really pertinent. But anyway, there are many, many constitutional law questions that come up and I think we’re dealing with so many issues at our school. 

We’re dealing with kids feeling alienated, alienated from the community, but also the work that they’re being asked to do every night. And it’s so interesting. There’s another book called Kids These Days by Malcolm Harris. It’s a wonderful book too, and he talks about if you think about what we’re asking kids to do, we’re asking kids to do tons and tons of work, hours of work that then quickly gets reviewed by the teacher. And maybe it’s not reviewed as you said, Richard, for months. And then it gets tossed and so they are literally wasting their time doing work that basically doesn’t do much good for anybody. And so I think they rightly feel alienated. And a lot of kids feel depressed. And they also feel like their work doesn’t have much meaning. And frankly, oftentimes it doesn’t. And so I think if we reimagined education, our kids were engaged in their community. It’s not just about the kids feeling engaged, but also the community felt like the kids were doing something valuable in the community.

There’s plenty of writing opportunities that you could be doing, different kinds of writing opportunities that you could be doing in this context. And you’d be learning about constitutional law. We could do that. And I think that chatGPT or whatever AI bot comes around next hopefully will force us into doing that more. I don’t think that, frankly, we will, and I don’t think that most schools have the capability to, because again, of the constraints of the tests that students have to take and the test prep. But that’s where I think we should be going. Because then you don’t really have to worry about chatGPT at all.

[00:10:25] Amy H-L: So you’re talking about community service, or really experiential learning, not community service work. 

[00:10:32] Lev M: I’m thinking about how you can reimagine what a good education is and what schools are for and what students ought to be doing with their time. There was an op-ed in the Times in the last couple years, during this period of Covid, where the person basically said, look, let’s be honest about what schools mostly are is that they are teachers are babysitters. There are a few schools which are doing really excellent stuff, but basically we need someplace to put the kids all day. And I think that the kids understand that. And again, I think if I’m thinking about a lot of my friends who were in other schools in the city, what they see, what they tell me about what’s going on in the classes is pretty depressing. Everybody knows the deal and we spend lots and lots of money on education in this country and in this state, in this city. There’s no reason why we couldn’t imagine the role of schools. And again, if the kids were doing meaningful work, you really wouldn’t have to worry about AI. 

[00:11:39] Jon M: So I’m wondering if you can sort of elaborate on that a little bit more, maybe give an example of your thoughts about Port Authority and people who are coming there. How do they tie together? 

[00:11:57] Lev M: Okay. I’ve already sort of talked about looking at it from a legal lens, but you could have a class studying those in an academic context, studying those issues, the rights of asylum seekers, for example, and then going in, doing interviews with people who are seeking asylum. You could have other people going to the shelters that the people are in and being journalists and writing about what they see at the shelters, interviewing the people who work at the shelters as well. One of our colleagues is also wonderful. She has a class on the history of New York City in the 1980s, and so she has every week guests come in who were living in New York, were activists in New York, were in city government in New York, were in finance in New York, and we’re talking about that period. She has someone come in who was a squatter in the 1980s and talked about the wars with slum lords. That’s great. And we could have kids going out into the Bronx, for example, and looking at what squatting looks like now, and doing interviews, writing oral histories, doing real world projects, meaningful projects, getting them out of the building.

When I was in, I’m sorry, I’m going on. Tell me, Richard, if you want to jump in. You were at Central Park East as well, but I remember so clearly Central Park East, my elementary school. It’s a school started by Debbie Meier in East Harlem. And I want to say it was one of the first. I don’t know if it’s the first progressive school, but it was a model progressive school. I remember our fifth grade class, a good part of the year we had to put out the school newspaper, and that’s basically all we did. And so some people were editors and some people were journalists. Some people were copy editors. Some people were in charge of actually doing the printing. Some people did cartoons. Other people did edits. That was our job. It was one of the roles that that a class had in the school. Other classes were in charge of making sure that from March to June that the garden functioned. There’s lots of learning that that can take place in that context.

[00:14:05] Jon M: So what I think you’re saying is that in that kind of setting, besides all the other advantages, that it just doesn’t become an issue of whether students are gonna just go to a chatbot to write something because it would just be totally out of context. And that if they did choose to, like, let’s say that they were looking at the constitutional issues of putting people on buses, that they would still have had to have the experience and then if they did, to access a chatbot for something about the history of migration, they would still have to tie it into real life. Is that…

[00:14:47] Lev M: Yeah. And why not? Right? We do it all the time with Wikipedia. I think the real challenge is if you’re asking them to write a formal academic paper, and I think that there’s space for that, I think you need to know how to do that. I think we do a little bit too much of it now in our school because we’re, it’s as if we’re training everybody to become PhDs and that’s not what most people are gonna want to do. But you do have to know how to write an essay. And I think to go back to Richard’s point, we need smaller classes, we need more teachers because we do need more time with the kids. 

But you could structure an essay, which I know. Richard did and I do. You could structure an essay. I’m doing one on the Haitian Revolution. It’s taken us six weeks. Now, again, if we had to prepare for the Regents exam, you couldn’t do this. But it’s a six week essay where you bring real books into the classroom and they hang in the classroom with us for a week and the kids open them up and they discover, and it sounds silly, but they discover the index. They’ve never seen this before. It’s this incredible search engine in the back of the book. And they have also index cards and they write on one side. We teach them how to take notes on an index card. And part of what they’re turning in every week are these piles of index cards. We’re helping them organize their research and for a big chunk of the six weeks, they’re not using computers at all. And then when they do go to the computer, we help them use JSTOR. They’ve never seen JSTOR before. We help them figure out what JSTOR is. 

[00:16:21] Richard M: For people who may not know, jSTOR is a search engine. It’s a catalog of academic articles, probably the leading place to find scholarly articles on almost any topic. And again, a place like Beacon has access to it. I think maybe other public schools can get access through…

[00:16:40] Lev M: They actually don’t anymore because like the budget stuff, but the New York Public Libraries made it accessible to everyone with a library card. So even when they have to go on JSTOR, which is on the computer, like they’re taking notes on index cards and they’re showing you that along the way, so you know, you’re scaffolding the project and they’re turning in notes all the time. And so it just wouldn’t make any sense to use AI to write your paper. It wouldn’t actually. 

[00:17:13] Richard M: Right. If I can just say in, in sort of in terms of the overview of research essays, that if you already are structuring the essay like Lev does and, and like I did, then I don’t think chatGPT is a big problem because students have to show work along the way and it’s much more about the process. And even before chatGPT, we had students who did no work along the way, and at the last second would come in with a full-blown paper, perhaps good, perhaps terrible. And they either had somehow written it themselves or cobbled it together from the internet or somebody else wrote it. So I think probably at a lot of schools where teachers don’t have the ability to track the process of what their students are doing, this could be more of a problem because all of a sudden the student shows up with a fairly well-written paper and they’ll get a B or a C on it, and the student will be quite happy and they will have done almost no work and not learn much about the process of research and thinking and changing their research question. and getting feedback from a teacher, again, because for a lot of teachers, they can’t give that feedback. So I don’t think it’s a game changer if you’ve already been doing a pretty good job of the process like most of the teachers at Beacon. I think probably in a lot of schools, this is going to be a game changer where kids will just be like, oh my gosh, this is great.

Even at a place like Beacon, we had lots of kids getting over on homework and small assignments. It happened a lot in my class and really quote/unquote good students turning in work that wasn’t theirs. Why? Because they were exhausted from sports practice or because something was going on in their world. I think that’s gonna happen an awful lot. So on a daily basis, I think this is gonna make teachers crazy and they’ll end up with their hair looking like Jon’s and mine because they won’t know what to do about the daily onslaught of, oh my gosh, this kid is getting over, this kid’s not getting over.

On the bigger picture, in a place like Beacon, I don’t think it’ll make a big difference. I think there might be some positives for schools where they’re not teaching writing very well right now. It might actually help kids who are struggling, who get no feedback from their teachers, who their teachers don’t know how to structure the essay. Again, I haven’t been able to get on chatGPT because they’re overwhelmed already. But Lev has. But all the articles I’ve read say they’re turning out decent outlines and decent essays, and that could be helpful. 

One other thing. From a teacher’s perspective, the ability to write exemplars and give students three examples. Here’s a good essay on the Haitian revolution, here’s a great essay, here’s a weak one. And I used to write these myself. To be able to just generate those and then bring them into class, that’s a great time saver for teachers. And I think the one positive for teachers could be some of the busy work that we currently do, checking homework. If the chatGPT can do some of that, it could give you more time to think about what’s a big question to ask the kids? What do I wanna do in class differently than I’ve been doing heretofore and not spending, and I spent a lot of time on this, checking homework every night and giving feedback on early drafts of essay. That could free the teacher up. One of the things that struck me was how much excitement there is among teachers about how this could be a game changer for them. Again, these may not be great teachers. Maybe they’re already kind of skating over. I don’t know. But it could help.

The last thing I would say, teachers already spend a lot of time trying to figure out if kids are cheating, unfortunately. We use for essays and all kinds of other filters. The horse is already out of the barn on kids using the internet. Lev and I have a famous example going back to, I think 2006, 2007, where a girl cobbled together an essay that was quite good. She hadn’t written a word of it, and she even had an original interview that she’d gotten off the internet somewhere else and put it in and told us that she had met the woman at the Starbucks up the street. And this is 15 years ago,.

[00:22:05] Lev M: Richard. I also, I think the woman that she interviewed was dead at that point.

[00:22:09] Richard M: Yeah, it was from the 1930s. It was a great interview. She had lived through the Depression and yeah, she was dead. Or not even real, who knows? Yeah. And she was adamant that she had written that project. Yeah. So this is not really new. This is just, it’s more difficult.

I do think it’s amazing how much drama there is around it right at this very moment. Like, oh my gosh, the college essay is finished, high school English is finished. I think that’s because a lot of teachers are not doing a very good job already. But again, I think because frankly a lot of teachers aren’t doing a good job, that this could present opportunities as well as real pitfalls. Yeah. Again, I’m glad I’m retired. 

[00:22:58] Jon M: Lev, you’ve been skeptical of the influence of technology on young people. Overall, what do you think the effect of this technology will be? You mention Neil Postman, for example. 

[00:23:09] Lev M: Neil Postman has, I don’t remember which of his books. Maybe there’s a book called Technopoly. And he says, basically, there is no technology that’s going to make a bad teacher suddenly a good teacher or a bad classroom, a good classroom. So if it’s a terrible classroom and you bring in, if everyone gets laptops, it’s not gonna be a good classroom. And so I, I think to Richard’s point, I think if you’re creative, when you’re doing exciting stuff, you can use this technology. But I’m not of the mind that technology is neutral. And I think, again, I get that from Postman or McLuhan. I think we’re gonna lose a lot with this technology. I don’t know what that is, though. 

I don’t know, and I don’t really want to speculate, but I think to go back to what Richard was saying about why now, why people are sort of freaking out now. I think it’s a couple things. One, I think that it, a lot of this is about understanding that many, many people’s jobs are, are going to disappear in the future. So there’s a general anxiety around what this means for work for adults, and maybe that also has something to do with the nature of so much of our work today being as David Graeber put it in a wonderful book, Bullshit Jobs. So many jobs are bullshit jobs and people are– sort of know that. I think he does a good job of documenting that in his book. And everybody should read the book. I just finished reading it, and so I think there’s that. I also , to go back to the beginning of the conversation, many teachers, most kids, I would say, and probably a lot of adults recognize that what kids are doing in school is not particularly relevant to anything that they’ll be doing later on. And so I think there’s this general feeling. You see a lot of this in the press. We’re gonna be replaced by AI. And again, it’s not the way I think it’s gonna turn out. I think it’s gonna turn out terribly because I’m predisposed to think that that’s the way things are gonna go. But it could actually be, to go back to Keynes, it could actually be this amazing thing where we do a whole lot less work and we hope have a whole lot more time to do whatever the heck we want to do. And that is what Keynes predicted about a hundred years ago for human beings. And you know, he sort of had this idea of what growth would be, GDP would be, and he was wrong. We’ve like doubled or tripled. We are so much richer than he could have imagined. And he thought we’d be working a few hours a week, and the opposite has happened. So this could be incredible. Again, I don’t think that’s what’s gonna happen cause that’s not been the direction of things. But maybe

[00:25:50] Richard M: if I can just add one more thing and I didn’t read the article, but someone wrote an article that Google already has a better chatGPT than chatGPT, but they haven’t unrolled it because I don’t know, proprietary issues or licensing issues, but I think that could be problematic because so many of the schools, including Beacon, have already gone over to Google Classrooms. And if Google has some kind of similar function, I could see this being rolled in as a way back to what Lev was just saying, to replace teachers to some extent. Because if you don’t need the teacher to check the homework and you don’t need the teacher to generate… Some teachers are talking about this is great, I, it’ll generate my quizzes for me. And it’s probably true. What are the five causes of the the downfall of the Ottoman Empire, or why is the Russian Revolution, someone asked me recently, important? This might be able to do a better job or quickly scan the internet and, and get you some kind of competent answer. So I think there are all kinds of big implications in terms of the money aspect of this for teaching that no one’s written about that much yet either.

[00:27:06] Lev M: Yeah. And if I could just I jump in, I think it’s certainly going to de-skill the profession and that could lead to the collapse of our union and the kind of protections that we have, and certainly our wages. But I also I think there’s the thing that’s sort of missing in the conversation, that I’ve been thinking about a lot. I was listening to a podcast with this guy in venture capital, and he was talking about substack. I don’t know if you guys are familiar with what substack is, but I can explain it. They’re basically, you subscribe to the writers that you like. A lot of them used to be for the Times or Washington Post or whatever, and they’ve gone and they’ve become free agents and they have their own newsletter that comes out, many of them daily. And you pay a little bit of money every year to get their thoughts in your inbox. And what he said, that sort of substack’s idea of what they’re looking for from their writers, when they pick writers to work for them, is non fungible ideas, nonfungible columns. And what I think he meant by that is, or they meant by that is, that sure AI is going to be able to produce a fairly good, fairly unexciting essay. And it’ll only get better with time. And they’re already, there’s already AI I just read about this, in Israel, which some Israeli company is gonna put out, which can already give you, which can cite sources. One of the things that chatGPT can’t do is cite sources. And we’re gonna get that very quickly. So it’s gonna get better. But what I don’t think it’s going to be able to do is… 

The other day I was messing around with ChapGPT and seeing if I could write a Haitian Revolution paper. And so I asked what are the three leading causes of the Haitian Revolution? And it gave me three good causes. But then what’s so cool about it is you can say, Give me this paper, but include the ideas of Laurent Dubois and C. L. R. James, who are two of the big historians of the Haitian Revolution, and it could do that, too. And it was so much better. In other words, you kind of have to know what to put in for you to get something interesting. And I think you’re, you’re always gonna need to do that with AI. And in that sense, what we need to be teaching kids is to know who, and I think Richard did this really well, but know who the big people are in the field, what the big debates are, and you can create interesting content. But you know, garbage in, garbage out and quality in, quality out. And we need to be helping the kids understand what is quality. 

[00:29:53] Richard M: You know, Lev, that it is kind of interesting because you could give the project, start with three major causes of the Haitian Revolution. Then you could say, who are the major historians of the Haitian Revolution? What are the major historiographical disputes of the Haitian revolution? Who’s the leading critique of C. L. R. James? So you’d have to teach them that, but once they figure that out, and maybe that’s again, for a high school student would be kind of amazing. Like in other words, if you could leave high school thinking through that process, particularly historiographically, maybe that’s a great thing.

[00:30:33] Lev M: Yeah. 

[00:30:34] Richard M: At some point, it does seem to me, they don’t need you or me to do that. In other words, if the program is that clever, you could say, okay, start with the three major causes. They write kind of a bland essay, and then it prompts you to say, well, okay, but there must be some historical dispute about that. What’s a Marxist perspective on that? What’s a liberal perspective?

[00:30:58] Lev M: I’m not worried about that, Richard. You know why? Because you can do that on Wikipedia already. You can go to historiography of the Haiti Revolution, but the kids don’t do it.

[00:31:05] Richard M: No, but this is even easier. 

[00:31:06] Lev M: No, it’s so easy already. The thing is the kids are, the kids would never go do that on their own, right? You have to be brought to the point where you…

[00:31:13] Richard M: Or you have to teach them how to do it. 

[00:31:15] Lev M: Absolutely. The other thing I was thinking about, Richard, when you were talking about cheating. You know, I actually don’t spend that much time thinking about cheating. I do when we think about these big papers, but basically, I give the kids homework at night, and the truth is that I’m reading their homework pretty quickly so I can prepare for class and see what interesting things kids said or where they were like as a whole, where they were not understanding things so that I can try to address those things in class. And so for that, homework is really, really helpful. But they basically get full credit for attempting the thing. 

I think there are two ways that you can make sure that there’s no bullshit. And I think, excuse me, again, I’m cursing a lot in this episode. I think one of the ways is you give kids small quizzes, small frequent, low stakes quizzes, which is the opposite of these big exams that we’re doing on a national level or state level. And you keep asking them same or similar questions over and over again throughout the semester. So you reinforce the stuff. And the truth is that there’s really not a way to be asked that. If you’re using AI for homework, you’re not gonna be able to do well in these quizzes. But again, you have to trust that the teachers will be able to make these quizzes.

And I think part of the problem that we have, Is we don’t think very highly of our teachers, which is why we have standardized tests in the first place and why we pay, for the most part in this country, our teachers very, very little. But I think if you started to, like, I know you guys had a show with Sam Abrams about Finland a couple years back, but if you start to have confidence in your teachers, pay them a bit better, relatively better. And I think that’s the idea, is you’re paying them in Finland, not millions of dollars, but you’re paying them like other professionals, you know, are paid in Finland, and you give them a month extra to plan. So the school year is shorter, but you give them an extra month to plan together. If you do all these things. I think, like we’re seeing in Finland, you’re gonna get creative people doing creative work. So that’s one way you can, again, not worry so much about AI and not really worry so much about cheating. 

And I think the other way that you do it is you have interesting class discussions with kids. Again, I don’t think this is happening in most schools, but you could do that. And through discussion you could figure out who’s doing the work and who’s not doing the work, which you and I sort of know anyway, Richard, right. 

[00:33:44] Richard M: I agree a hundred percent with that. I think in some ways, for students who are struggling or who are lower skilled in terms of both literacy or just comprehension, again, I haven’t thought about this very much, but this could help them. They could go home at night, get some ideas about how to engage in the conversation. Because oftentimes lower skilled students come to class and they get overpowered by the kids who quickly have figured it out or who are talking in a way that they can’t mirror, and they then clam up. But if somehow they’ve gotten sort of a headstart with some of the basic concepts, maybe this could get them rolling. I have no idea, but I know that with a lot of the special ed kids, that their teachers would, say, help them to pose a question at the beginning of class or get started. This might be a help.

I just wanna say in terms of checking kids who are gonna try to get over, I think we have to be honest, kids are gonna try to get over. Come on. And even I had lots of examples of really good students who just couldn’t do the homework that night, and you know who I’m talking about, Lev. And I caught them just by luck with checking their– because I was speed reading their homework. And they were very honest about saying, look, I just couldn’t get to it tonight. I was overwhelmed. I had a game, I had this, and that’s gonna happen a lot with chatGPT unless you change the kind of homework assignments. And so you’ve gotta make it somehow impregnable to the chatGPT, whether it’s some kind of personal connection, but otherwise it’ll be used rampantly. And we know this already from language classes, where they’ve totally gotten rid of homework because the kids were just using Google to do the translation, and so they don’t give homework anymore. 

[00:35:41] Lev M: Right Richard. But that could be a good thing if we don’t give homework anymore.

[00:35:44] Richard M: Well, it could be if, unless you’re Lev Moscow.

[00:35:47] Lev M: But I’m glad you bring up the example of language classes because, not at our school, our school’s pretty great with it, but in general, language classes are like the worst classes. They’re doing it all wrong. And so again, my point earlier on was if you did language classes well and kids were actually having to use language, interact with people. Let’s go back to the asylum seeker stuff that we were talking about earlier. Let’s say you had a language class where you’re going, one of the things the volunteers need are more people who can speak Spanish. I don’t speak Spanish well at all, and so I’m not much use, but they need people who can speak Spanish because almost everybody coming in now is from a Spanish speaking country. And what you do, what they’re doing at at Port Authority is you are basically taking orders of what people need. They need two pairs of jeans, one pair of sneakers, and a sweater. And you have to be able to speak to people in Spanish. What if you brought your Spanish class there and you spent two hours talking to real people about their experiences and you’re helping people out and you’re being useful. Of course people are using people are using Google Translate in Spanish classes now or French classes cause they’re terribly designed. 

[00:36:53] Jon M: It’s very superficial compared to what you’ve been talking about, but I suspect it may be a question that is still hanging out in some people’s minds, some listeners’ minds. You mentioned at the very beginning, Lev, that one of your colleagues had already caught a couple of kids who were using chatGPT, and I’m just curious if you happen to know how they caught it, because I know some of the articles and stuff that I’ve been reading have been talking about because it isn’t exactly plagiarism, because it’s not an article that somebody else has already written or a paper that someone else has already written, that some of the traditional software that Richard mentioned, for example, doesn’t catch stuff. I’m just curious if you happen to know how they happened to notice that it wasn’t theirs.

[00:37:38] Lev M: What they said was, for this particular assignment there needed to be argumentative topic sentences. And chatGTP right now, although maybe it’s changed the last week, isn’t very good at that. I’m sure it will be someday, but it’s not good at that yet. There were no, again, no citations. It doesn’t know how to do that yet. And so this assignment asked for in-text citations, which it wasn’t able to do. And then in this case, I don’t think that the kid was amazing writer, and the quality was a little bit better than what the kid had been producing all year. And so it just seemed kind of off. So that’s how they caught it. I understand that there’s some kind of watermark, like virtual digital watermark that is created when AI produces an answer and that chatGPT or OpenAI, which is, I guess, the nonprofit that’s develping this technology, they are making that available so that you can use a program to run and you can run a program to see that watermark and see that it’s been produced by AI. So that already exists if you wanted to use technology to unmask the technology.

[00:38:47] Richard M: Right. The other thing it’s not just technology that’s a problem with high school writing. So we’ve had lots of examples of students having outside tutors basically do their writing and that’s a whole issue of equity in terms of kids with means. But that happens a lot. And students getting all kinds of help, not from chatGPT, but from an outside tutor or their parents, which creates all kinds of disparities. And so you have students that you know from what they do in class, are struggling to put together thoughtful ideas, and then they come in with a finished product that’s way beyond what would seem to be in their capability. And that has to do with process. And again, and I think Lev does a great job of this. So if you track what kids are doing all along from the beginning of their ideas, the questions they’re asking, annotated bibliographies, writing part of the essay in class, a first draft, all of that can stop a lot of the problems that existed way before chatGPT, with just the internet or going to the local library and copying out of an encyclopedia. Like I did back in high school.

So there are real ways to stop this, all of which has to do with the teacher’s ability to structure the class and the willingness to put in an awful lot of work. Yeah. And chatGPT can help with some of the work in future, but if they’re not willing to do that, then the kids are gonna be cheating and not, I don’t wanna say cheating. They’re just gonna be getting over because they’re already getting over in some other way. This will just make it easier and their essays will be slightly better. 

[00:40:31] Lev M: Richard, I totally agree with you. I just wanna add one more thing, which I think you were the master at, which is not just structuring the assignments so that kids can’t cheat. It’s creating an environment in the classroom where the kids care about… What was it that Jack Marooney wrote about, that he knew more than Eric Foner? What was that? 

[00:40:53] Richard M: West Virginia secession, . 

[00:40:54] Lev M: And this kid was really excited when Eric Foner came to your classroom to talk to the kids. And Jack, who had written his senior paper about West Virginia secession, was, and I probably still remember this moment, he was really excited that Eric Foner said, “Look, kid, I think you know more about this than I do.” That’s because you created that in the classroom. Again, I think it’s really important to create tight assignments that are kind of tamperproof and you can’t cheat, but it really comes from your knowledge, your passion, the ability for you to be creative at your job, and you are a singular talent in a lot of ways, but so many people who are just as smart or talented as you are not in teaching because I don’t think currently they feel valued. And I think we also want to be honest, and you’ve told me this before, that the reason you can’t create very many good schools is because you need a lot of teachers who will spend a lot of time, and talented teachers who will spend a lot of time in those schools working, and that can’t exist given the current climate, political climate. And also, yeah, the way we, we treat educators and the value we have collectively for education. 

I’ve been thinking a lot this year, and this may sound silly, but I think it relates, I’ve been thinking a lot this year. You were talking, Richard, about lower skilled students. I have a student who’s, I wouldn’t even say, lower skilled. He’s just like not that interested in what we’re teaching in class, but he’s really into the NBA more than any student that I’ve ever had. And I am really into the NBA, too. And so we talk a lot about the NBA. And I was listening to a podcast, one of these like really nerdy NBA podcasts, where these guys talk for two hours about stuff that the casual fan’s not interested in. It was like a two hour podcast where they ranked.

[00:42:54] Richard M: You can only do this now because you got chatGPT correcting all your homework.

[00:42:58] Lev M: Exactly. So you have these guys talking about ranking the 30 best NBA franchises based on the last 10 or 15 years, based on how they do with player development, with drafting, with free agency, hiring coaches and so forth. This is super interesting to me. So I’ll say to the kid, listen to this podcast, and then what I want you to do for extra credit this semester is write for me these 30 paragraphs where you rank the teams, and you have to give me data and so forth, and explain why you’re ranking the teams. And this kid’s been working on this since October and it’s amazing. It’s the best thing that anyone is producing in any of my classes. And I keep thinking about this, like, why should I be reinforcing how important the Reformation is, when I am the only one in the room who really thinks the Reformation is interesting at all. But the Reformation is when the ABA split and then starts the NBA, it’s when they merged. It’s when the two leagues merged. You’re making some connections for the kid. So I was thinking again, so much of what we’re doing is clearly meaningless to the kids. And if things are meaningless, then you you cheat. And you use, you’d use AI. So let’s figure out a way to make things more meaningful. Yeah.

[00:44:28] Jon M: On that note, thank you, Lev Moscow and Richard Miller. Thank you very much.

[00:44:34] Lev M: Thanks for having us. Thank you guys. 

[00:44:37] Amy H-L: And thank you, listeners. Check out our new video series, What Would YOU Do, a collaboration with Dr. Meira Levinson at the Harvard Graduate School of Education and EdEthics. Go to our website,, and click video. In the first case study, a teacher using Action Civics faces pushback from a parent. The goal of the series is not to provide right answers, but to illustrate a variety of ethical viewpoints.

If you find this podcast worthwhile, please share it with a friend or two. Subscribe wherever you get your podcast and give us a rating review. This helps others to find the show. Check out our website for more episodes and articles, and to subscribe to our monthly emails.

We post annotated transcripts of our interviews to make them easy to use in workshops or classes. We work and with consultants to offer customized SEL programs, with a focus on ethics, for schools and youth programs in the New York City and San Francisco Bay areas. Contact us at We’re on Facebook, Instagram, and Twitter @ethicalschools.

Our editor and social media manager is Amanda Denti. Until next week.

Click here to listen to this episode.