Episode 137: Transcript

Episode: 137: The Creativity Fallacy

Transcription by Keffy

Annalee: ​ [00:00:00] Charlie Jane, I need to tell you about a scientific paper that is really annoying me. 

Charlie Jane: [00:00:07] Science gossip is life. Lay it on me. 

Annalee: [00:00:10] Okay, so it came out recently in Nature Scientific Reports, and it's called, quote, “Best humans still outperform artificial intelligence in a creative divergent thinking task.” You have to love a scientific paper that has best humans in its title, because after all, best is certainly quantifiable.

Charlie Jane: [00:00:31] Also, was that headline written by AI? I just pictured AI being like, best humans, worst humans, like? 

Annalee: [00:00:40] You have to wonder. So, I'm going to tell you why this was bugging me so much, because, not so much this paper specifically, but it kind of represents a bunch of different studies that are being done on AI right now. And it's just kind of a perfect example of how AI is really mangling our idea of what creativity is. 

[00:00:59] So, there's these two cognitive scientists who wanted to know if humans could be more creative than large language models based on different versions of GPT, which is the model that fuels, ChatGPT. So, they found, that, on average, GPT was more creative. So, the supposedly “best” humans in the title, they were more creative, but on average, humans were less creative. So, the part that really pisses me off is how they measured creativity. 

Charlie Jane: [00:01:31] Oh god, I'm bracing myself. 

Annalee: [00:01:34] So, they paid people a few bucks to do what's called a divergent thinking task using, I assume, Mechanical Turk or some other gig app. They got people, they found them online and they paid them a few bucks and here's what they had them do. So, these people were given four words. The words were, candle, rope, box, and pencil, and they had 30 seconds to come up with the most creative uses for each item. 

[00:02:02] So, basically the word candle flashes on the screen, and they have 30 seconds to come up with, I don't know, bad dildo, mouse bonfire, fart ignition, whatever.

[00:02:12] And then the GPT models were given the same task. But then, guess how they evaluated creativity? You have 30 seconds to come up with something creative. 

Charlie Jane: [00:02:22] Oh my god, I’m just, I'm already cringing. I just… I assume that it's going to be something super mechanistic. 

Annalee: [00:02:29] So, what they did was use this measure called semantic distance between the item, like, say candle, and the creative use. Semantic distance is a term that measures how far the meaning of one word is from another. So, like, river and stream have a small semantic distance between them. But river and metaphysics have a really big one. So, you can chart this out on nodes on a graph. And they also decide that semantic distance, like, the greater the distance, equals creativity. So, the further apart those two words are. And then they also randomly seemingly brought in a panel of, they said, lightly trained humans to measure the creativity of each answer on a scale of one to five. 

Charlie Jane: [00:03:18] Lightly trained. 

Annalee: [00:03:19] Lightly trained. They actually…

Charlie Jane: [00:03:23] Lightly trained. I think my FetLife profile mentions that. Lightly trained. Anyway. 

Annalee: [00:03:28] Yeah, so all of this sounds really fucking arbitrary to me as a measure of creativity, especially when you take into account that not everyone can come up with creative shit in 30 seconds for a few bucks online. 

Charlie Jane: [00:03:42] Well, and creativity doesn't exist in a vacuum. Necessity is the frickin’ mother of invention. Like, if you give me a candle, it's like, well... what's my scenario here? You don't just be creative without any context. 

Annalee: [00:03:52] And finally, okay, so this is what really, really, really pissed me off, is that they discounted any answer from a human that they deemed illogical or confused. And I mean, isn't one person's illogical, confused, answer another person's creativity?

Charlie Jane: [00:04:14] Probably the people who were illogical were just less articulate or didn't talk in a middle-class accent. I'm just going to guess. This is giving me flashbacks to horrible IQ tests. I mean, it's just, it's ridiculous. 

Annalee: [00:04:26] Yeah, it's really ridiculous. And this is the kind of weird bullshit that has led me to this moment.

[00:04:32] You're listening to Our Opinions Are Correct, a podcast about science fiction, science, and whatever other creative words you can come up with in 30 seconds. I'm Annalee Newitz, a science journalist who writes science fiction. My latest novel is The Terraformers. 

Charlie Jane: [00:04:47] I'm Charlie Jane Anders. I'm a science fiction writer who knows a little bit about science and my latest novel is Promises Stronger Than Darkness.

Annalee: [00:04:56] So today we're tackling the myth of creativity. We'll be talking about how AI is commodifying creativity in a whole new way. But we'll also be getting deeper into questions about what it means to create. We're going to talk about the status of the author, and how we've idealized artists at some points in history, while at other points we've celebrated the so-called death of the author.

[00:05:19] Plus, you get to learn about the intentional fallacy. 

[00:05:26] Later in this episode, we're going to be joined by friend of the show, Mary Anne Mohanraj, an author and professor of literature at the University of Illinois at Chicago, and she has a lot of interesting thoughts about using AI in the college classroom.

[00:05:39] Also, on our mini episode next week, we'll be talking about some cool shit you should know about octopus creativity. 

Charlie Jane: [00:05:46] And that reminds me, did you know that this podcast is entirely independent? We're not part of the Blarg stream or the Wiffle network or any of that stuff. 

Annalee: [00:05:57] So creative!

Charlie Jane: [00:05:57] We're just like a lonely little podcast sitting there on an island beaming our words out into the void and hoping that people will find us. And that means we depend on you, our listeners, for your support and your love and just your community. And the way you become part of our community, really, is by supporting us on Patreon, where even just a few bucks a month help to keep this podcast going. And you get to join our discord channel where we hang out and gossip about science like nonstop. 

[00:06:28] So, you know, just think about it. Whatever you can afford to give us, whether it's a few bucks a month or 20 bucks a month, everything goes right back into making our opinions that much more correct. Find us at patreon.com/ouropinionsarecorrect. 

Annalee: [00:06:41] All right, let's get creative.

[00:06:45] [OOAC theme plays. Science fictiony synth noises over an energetic, jazzy drum line.]

Annalee: [00:07:17] I've been thinking a lot about what it means to be a creator, and that's partly because the writers strike in Hollywood, and partly because I feel like we've just been inundated with the studies like the one we just talked about. 

Charlie Jane: [00:07:28] Yeah, it's a really dark time. Even before AI came along, creative people have been getting squeezed for a long time now.

Annalee: [00:07:34] Yeah, and one of the big issues in the writer’s strike is that the Writers Guild is asking that authorship on scripts not be assigned to AI. And there's a really simple economic reason for that, which is that if an AI is given author credit, then the studio doesn't have to pay anyone. Writers in the Guild can only move up the pay scale if they get writing credits, so it really matters to their livelihood.

[00:07:58] And luckily, or interestingly, I should say, in August, a judge ruled that AI-written material can't be copyrighted, so that actually helps the writers’ cause here. But that's a really economic definition of creativity, right? Being creative means being credited on a script and getting paid. 

Charlie Jane: [00:08:19] Like I said, necessity is kind of the motherfucker here and part of creativity is that you get really creative when you need to get paid.

Annalee: [00:08:27] Yeah, that is really true. And there's another way that creativity is being measured, which also touches on AI. Which is this notion of originality. And this kind of goes back to what we were talking about with copyright.

[00:08:39] So, originality basically just means coming up with something that feels new. And there's a number of writers and artists who are suing companies like OpenAI and Meta for using their writing and their art as training materials for algorithms like ChatGPT and MidJourney. And other artists and writers are organizing protests online against the use of these algorithms.

[00:09:07] And so, basically what these creators are claiming is that these models are stealing their ideas and regurgitating them in slightly different form. Of course, the AI makers would say, oh, these people don't understand how AI works. That's not at all what's going on. But the writers and artists are saying, humans are the truly creative ones, and that AI companies are actually breaking copyright laws to make their models appear to be as creative as people.

Charlie Jane: [00:09:34] Yeah, you know, it's funny, I used to really hate this idea that there was no such thing as originality, that every story had already been told, and that we're all just like, remixing Homer and Shakespeare and whatever. 

Annalee: [00:09:44] Yeah. 

Charlie Jane: [00:09:44] Which I think it’s both true and not true. Like I think it's possible to come up with stories that are like new in the sense that they take you to a new place or that they kind of have a new spin on the old tropes.

[00:09:57] But I also think that You know, as I've gotten older and mellowed, I've sort of come around to feeling like yeah, I mean, I feel like we're all just like floating in a sea of ideas and we're all kind of just like passing them through our membranes and like maybe we like re-jigger them a little bit as they pass through and get osmosed and then poop out again. I don't know. 

Annalee: [00:10:15] But I think there is something to be said for the idea that humans can take an idea that might be recognizable and have it respond to new historical circumstances and new cultural circumstances in a way that you might not see an AI doing. And so, at any rate, that's the argument in these lawsuits is that they would kind of disagree with you. I think they would say like, yeah, we live in a sea of ideas, but only humans really can tweak them and make them original. 

[00:10:43] So, we've got these legal and economic debates over creativity, but what does it actually mean to be creative? 

[00:10:51] So, Charlie, I know you're used to this happening all the time off the podcast, I want to info dump to you. And I want to take you on a not at all comprehensive journey through the last 120 years of academic approaches to what it means to be an author. Are you ready?

Charlie Jane: [00:11:07] I've changed my mind. Science gossip is not the best gossip. Lit theory gossip is the best gossip. Lay it on me. 

Annalee: [00:11:16] Well, I'm going to try to make it as gossipy as possible, but it may just be kind of like philosophical disagreement levels of gossip.

[00:11:21] So, one thing that I always find mind blowing is that in the United States, we didn't really have English departments in universities or even really any modern language departments like French or German until the late 19th and early 20th centuries. So, there's this whole professional class of literary scholars who start defining what it means to be an author right around the turn of the 20th century.

[00:11:48] Now, before that, of course, you had literary studies and literary criticism, but mostly they focused on the classics, like things written in Greek and Latin. And so, these early literary critics who were writing at kind of the dawn of English departments are mostly getting their ideas from Classics departments. And in those places, authorship conferred a certain amount of authority and your goal as a reader was to understand what the author meant and to sort of bask in his brilliance because of course he was a he, even though most of the popular novels in the 18th and 19th century were written by women, right?

Charlie Jane: [00:12:25] Right. I mean, I've been doing a deep dive on 18th century literature. There are all these like incredible women authors who were systematically memory-holed in the 19th century. It's kind of bizarre.

Annalee: [00:12:36] Even though they sold a ton of books. 

[00:12:41] So anyway, people are going around. At this point, they're reading poems and plays and novels and they're feeling really excited about just being allowed to teach that English, the language English, is just as literary as Latin and Greek.

[00:12:56] But then things start heating up. So, in 1946, right after World War II, these two American literary critics named Wimsatt and Beardsley wrote this essay called “The Intentional Fallacy,” which launched this whole new movement called The New Criticism. And their idea, which now sounds really old at this point, was that the author's intent was meaningless.

[00:13:25] So that's the intentional fallacy. And the way that people should read literature, and especially poetry, was that they should get meaning out of the text itself and just ignore everything that the author says about it.

Charlie Jane: Il n’y a pas de hors-texte! 

Annalee: [00:13:44] What? 

Charlie Jane: [00:13:45] That's French for, “There is nothing outside the text.”

Annalee: [00:13:50] And I mean, it's funny because you could imagine there's a certain economic benefit to Wimsatt and Beardsley, too, because, of course, if the intentions of the author are meaningless, you definitely need critics like them around on the payroll to help you understand the text, right?

[00:14:07] So, anyway, things kind of boop along, and the new criticism really takes over in English departments for kind of the reasons I was describing. Like, it really makes the critics seem super important. And then in the late, 1960s, this very famous French critic and semiotician named Roland Barthes, he kind of puts the nail in the coffin of the author with an essay, very famous essay called, “The Death of the Author.”

[00:14:33] And he's kind of agreeing with the new critics, like he's kind of rehashing their ideas a little bit, saying like, you can't really take the author seriously when it comes to what their work means. But he's taking it a little bit further. And I would say the death of the author essay really should be called the death of literature because Barthes argues there and in other places that anything can be a text and you can get as much social and cultural meaning from analyzing a wrestling match or a car or a popular movie as you can from reading great works of literature like Shakespeare. And his book, Mythologies, which came out in the late 1950s and also just like galvanized literary critics, it's this amazing collection of essays where he treats everything from like cruises to plastic as texts and shows how they're just as fascinating as great writing.

[00:15:26] So, you don't need an author to be fancy, you can have a mechanic or a wrestler essentially be an author. 

Charlie Jane: [00:15:35] I mean, I think that's beautiful because I hate this idea that author is a class of people, but I also think that you, when you reduce that to its like ultimate form, it's just like. You could literally just throw those magnetic things that you used to put on fridges up in the air and the words would land in any particular order, and that would be as meaningful as something that somebody spent an hour thinking about how to say a thing that was really hard to say.

[00:16:00] So, I don't know, kind of love it, kind of hate it. 

Annalee: [00:16:04] Yeah, and actually the fridge magnet thing is a great analogy because that kind of anticipates what ends up happening with, Barthes' ideas. So Barthes kicks off this obsession with semiotics among literary scholars, and it leads to a bunch of stuff in the ‘60s and ‘70s that I'm not going to talk about, like deconstruction and post-structuralism and feminist and Marxist theory, which I love, but all I want to say is that those are all kind of in the same vein of the idea that the author is dead, everything is a text, and, this is the key part, texts themselves are ambiguous and meaning is arbitrary.

[00:16:42] So like, everything is dead! Meaning is dead! 

Charlie Jane: [00:16:44] Nothing means anything! Nothing means anything!

Annalee: [00:16:46] So in a weird way, the poststructuralists are kind of anticipating AI creativity, because they think of meaning as arbitrary, and they view texts as existing in relation to each other, as opposed to authors being in relation to each other.

[00:17:04] So, you see texts referring to each other, you have these analyses of how, like, one text kind of speaks to another text, and in a way, what happens is, and I'm simplifying a lot here, but what's happening is that by the time the 1980s roll around, the author isn't just dead, she's actually become another text who exists in relation to a bunch of other texts, because remember, everything can be read, everything can be analyzed, and in a sense, nothing matters.

Charlie Jane: [00:17:37] I'm getting very tired. I’m getting tired. 

Annalee: [00:17:40] I'm sorry. 

Charlie Jane: [00:17:42] No, it's just like, it's not you. It's like, this is, wow. 

Annalee: [00:17:47] Like, I mean…

Charlie Jane: [00:17:48] Flashbacks to college, actually. 

Annalee: [00:17:49] Yeah, this is like the history of Annalee’s experience in a literature department in grad school. So, in the ‘70s and ‘80s, while you have a lot of this post-structuralism and deconstruction kind of coming out and really revolutionizing the way people think, you also get this other really interesting trend in how literary critics look at authors, or don't look at authors, really, and it's called reader response criticism. And this genre of criticism explores the idea that the way readers interpret the text is actually more important than what the author meant.

[00:18:25] So, like, we're way beyond the intentional fallacy. We're way beyond the death of the author. And now we're like, well, but who's alive? And the people who are alive are the readers. And a lot of social scientists, actually, like myself, when I was an academic, really liked reader response criticism because it lends itself really well to analyzing the way culture and art affects large numbers of people. So, you can use this kind of criticism to do things like ask the question, did the Barbie movie make more people think that feminism is good? And you would just go out and you'd interview a bunch of people to find out the answer. 

Barbie Clip: [Different Barbies and a Ken greeting the protagonist Barbie in the film.]

Hi Barbie! 

Hi Barbie! 

Hi Barbie!

Hi Barbie! 

Hi Barbie! 

Hi Barbie! 

Hi Barbie! 

Annalee: [00:19:09] The one thing you wouldn't do is ask the director, Greta Gerwig, what she thinks Barbie did. Because remember, she's the author/director, she's not the reader so nobody cares about her. 

Charlie Jane: [00:19:21] Yeah… I mean, I care. Greta, I care about you.

Annalee: [00:19:27] Yeah, we love you, Greta. So anyway, I think reader response criticism is in some ways the academic anticipation of fanfic. It privileges readers over writers. And I think that this question of whether the author's intentions matter, or the text matters, or the reader matters, those are still haunting the question of creativity right now. We're still kind of wrestling with these, both in literary criticism, but also just in everyday life and how we experience movies and books.

[00:19:59] So where do you think all of this stuff leaves the status of the author today? 

Charlie Jane: [00:20:05] Yeah, I mean, this is really hitting a nerve for me because when I first started trying to write fiction, like, professionally, 20 years ago, it was really… Like, the internet really kind of like banged into my skull that as the author of a piece of work, I was not allowed to comment on what it meant. I wasn't allowed to say what I intended. If I did tell people what I was thinking about when I wrote it, or if I did talk about my thought process or my intentions that I was robbing the reader of something like incredibly valuable. Because I was like trying to tell them how to read the text and that that was bad and that I should shut up and like, this wasn't directed at me personally. This was authors in general. It was a thing that people on the internet were really intense about in, I guess, the early 2000s and going into the late 2000s. 

[00:20:53] And it was just, it was this weird thing where it was like, yeah, I mean, authors, I mean, not always… Sometimes you're just like, oh, that seems like a fun idea. But authors do sometimes think a lot about what they're doing while they're doing it. And that's actually interesting to know about, I think. But on the other side, I feel like, yes, when you have a work of fiction, it is like a movie being played in someone's head and everybody's going to see a different movie because they're using a different kind of projector, I guess? I don’t know.

Annalee: [00:21:21] Yeah, and their theater is really different, right? Like if you go to an open air theater in the forest that's really different from like a gritty urban theater where like people are jacking off in the seat next to you. Not that that ever happens in real life.

Charlie Jane: [00:21:34] No.

Annalee: [00:21:36] No. I mean, I'm kind of torn because I remember when I first encountered this idea of, specifically, the idea of everything being a text, I got really excited because I was like, yeah, this is something that I have instinctively always thought and I always was like, well, yeah, like books are great, literature is great, but like also so is a cheesy television show or a trip to a really weird amusement park or something like that. 

[00:22:13] I felt like it was nice to be able to decenter the literary canon, or just even, like, the book author, like, to just be like, actually, yep, you're cool, but you know who else is cool? The guy who built that amazing bridge is really cool, too. Or not the guy, but, like, the group of people who built that amazing bridge and conceived of that amazing bridge.

[00:22:38] So I kind of felt liberated by that, even though I've always been a writer and I, you know, obviously I have opinions about my own work, but I actually really like the idea of like letting the work go out into the world and just like letting people do shit with it.

[00:22:51] And I mean, I also am happy to tell people what I think it's about, but I've had people say like, I don't think that that's what your book is about. And I'm like, sure, maybe it's not. I don't know. I don't know what's going on in my head unconsciously. And also, it's not my book anymore. It belongs to other people.

Charlie Jane: [00:23:10] Yeah, and I think that I kind of resist either extreme. I think that both versions are really good. I think that readers should be encouraged to have their own readings of things. I don't think there's an authoritative reading. I don't think that anybody, including the author, gets to like say, this is what this book means, or this is what this book says.

Annalee: [00:23:28] Yeah. 

Charlie Jane: [00:23:28] And like, it's now frozen in amber. I think that every reader gets to bring whatever they want to the book and gets to see it however they want and, you know, they could even skip like 50 pages in the middle and be like, well, I thought the book was about this because I didn't read that boring part.

[00:23:42] And like, that's legitimate. Do what you want, you know, you do you. But I also think that it makes me sad that we're going to go to the extreme of, authors should just sit in the corner and think about what they've done, kind of. Or not think about what they've done, because if they think about what they've done, they might talk about what they've done.

Annalee: [00:23:59] Yeah, no, I definitely disagree with that. And I think that there's a whole genre of like writing workshop not to get too far into the weeds about like what it means to be a writer. But like often writers go to these fancy writers workshops to learn about writing. And there's like one school of thought where, When the, so the author presents their work, and then everyone in the class says what they think it is, and the author's not allowed to talk, and you just have to kind of sit there. And like, literally, a lot of times I've heard horror stories about people just like, crying, like, people are saying such mean things, and they're just weeping, and they can't say anything, and they can’t—

Charlie Jane: [00:24:33] It’s so fucked up.

Annalee: [00:24:33] Defend themselves.

Charlie Jane: [00:24:37] Fucking Milford.

Annalee: [00:24:37] That’s called the Milford technique or something, right? 

Charlie Jane: [00:24:40] Yeah. The Milford method.

Annalee: [00:24:41] The Milford method. And a lot of writing workshops now are really pushing back against that. And they're like, no, we don't like that anymore, but still, that's like the extreme example.

[00:24:49] And I think that's just ridiculous. Like writers, this is the thing is like writers should be allowed to say whatever they want about their work, too. They're also readers of their own work, so…

[00:25:01] I wanted to end by asking you kind of like, one of those questions that really has no answer, which is what happens if you write a book and it becomes for whatever reason, it's not your fault, it becomes really popular with Nazis and like Nazis are like, oh my God, this sci-fi book proves that an ethno state will rise. And like, somehow, I've decided that your aliens represent my people and they're super excited about it. Like, how do you deal with that as the author? Because that's, that's kind of like reader response. Like some readers have found your book and they like it. 

Charlie Jane: [00:25:40] I feel like that's happened in real life, hasn't it? I feel like I can't think of specific examples right now, but I've definitely seen cases where that's happened. I feel like Stephen King gets that sometimes where he is just like, no, you misunderstood my book. Fuck off. 

[00:25:54] And like, I think that is what you have to do. Like I think that readers can make of your book whatever they want to, but if people come up with a really abhorrent reading of your book and really just like want to use it to justify horrible, horrible shit, then you do kind of have a duty to step in and be like, no, go fuck yourselves. You don't understand anything. 

Annalee: [00:26:12] Yeah, I think that the famous example is when Reagan used, Bruce Springsteen's song “Born in the USA” when he was campaigning. And Bruce Springsteen was like, no, that’s not it.

Charlie Jane: [00:26:26] Happens all the time with music, actually. Yeah. 

Annalee: [00:26:29] Yeah. And it's interesting because it touches a little bit on what we were dealing with in our Silicon Valley versus science fiction series where an idea gets taken up and used by a company or a person to justify their bullshit, and then the original author is like, I meant the opposite. 

Charlie Jane: [00:26:51] Yeah, I mean. This is getting in the weeds, but I'll just say really quickly I think that a lot of our most popular stories are ones where you can interpret it from like any angle like you could read Star Wars as being like a pro-USA story You can read it as an anti-USA story. You can read it as right wing You can read it as left wing and like that's why Star Wars is so beloved because it actively encourages every possible reading of the text. Its incoherence is why people love it. Like if Star Wars actually, successfully articulated, and I'm not talking about Andor, which is kind of more specific, but most Star Wars, if it actively expressed a particular point of view in a forceful or coherent way, it would be not nearly as popular as it is.

Annalee: [00:27:36] Yes, the Force would no longer be so forceful. 

Charlie Jane: [00:27:38] Exactly. 

Annalee: [00:27:39] That's a really great point to end on. And when we come back, we're going to talk to Mary Anne Mohanraj, who teaches literature at the University of Illinois at Chicago and has many thoughts about how AI can fit into student creativity.

[00:27:52] OOAC session break music, a quick little synth bwoop bwoo.

Annalee: [00:27:58] Joining us is Mary Anne Mohanraj, who is the author of several sci fi stories, including the collection The Stars Change. She also writes cookbooks, including her cookbook of Sri Lankan cooking, A Taste of Serendib, which she came on to the show before to talk about. And her latest book is a cancer memoir called Tornado. So, check that out. 

[00:28:19] Welcome back to the show, Mary Anne. 

Mary Anne: [00:28:19] So happy to be here. Hi, guys. 

Annalee: [00:28:22] Yeah, thank you for coming because we've been talking about creativity and how AI is sort of redefining it. And when I was in Chicago a couple months ago, you and I were chatting at Women and Children First, an awesome bookstore, and someone asked us about AI.

[00:28:37] And I gave this kind of standard answer of like, AI interferes with creativity, but you're a professor who teaches writing and you talked about how generative AI could actually be assistive technology for students and you had a much more positive reaction. So I wonder if you could just talk about that. Like, what does that mean for students to be using it? 

Mary Anne: [00:28:56] So I feel like I have to frame all of this by saying there's like no consensus in academia about this yet. So, everyone is trying to figure out what they think about it, trying to see what the effects are on our students. It's very early. We don't really have data yet on… students have been using it since last spring, or maybe even last fall, but the faculty may or may not have even been aware that students were using it and we don't have any way of telling what the long term effects are yet. Right? 

Annalee: [00:29:33] Right. 

Mary Anne: [00:29:33] So there's a lot of concern. I mean, I want to be clear. A lot of faculty are very concerned about this. You saw kind of an initial response that was very reactive, I guess, is, I would say where people were like, well, you know everyone's going to be cheating now. And so, in order to be disciplined and catch all the cheaters, which is not a mindset that I want to be in when I'm interacting with my students, but a lot of teachers kind of fall into that understandably because they get very frustrated by plagiarism and cheating.

[00:30:07] And so in order to do that, they were like, well, I'm just going to go back to like blue book exams in the classroom, and then they can't use the AI and as these discussions have been evolving on academic forums, people have been pointing out that's really problematic in terms of ableism, right? Making a student handwrite in blue books in the classroom, is problematic in like three different ways, right? 

[00:30:36] So, I don't think that's going to happen. I guess, you know, some people are trying to cope with it by can I come up with a creative enough assignment that it's hard for the AI to have a response to? That's tricky. 

[00:30:53] Can I do a lot of scaffolding work in the classroom? And a lot of us do that already. So that is kind of, if I'm going to have students who are writing a paper. First, they're reading the stories. I'm teaching an intro to science fiction literature class right now, right? So they read the stories. They keep a reading journal. We discuss in class things from their reading journal. Then, the reading journal is just their responses, right? So it would be, I think, tricky to have AI write anything coherent about their personal response to the story.

Annalee: [00:31:25] Yeah.

Mary Anne: [00:31:25] But then going on from there, the reading journals are meant to be notes so that when they're start thinking about writing a paper… We do that in the classroom too, where we try and come up in small groups. We brainstorm, come up with a thesis, then we draft outlines on another day, then another day we're working on rough drafts. And so there is, kind of, by the nature of that kind of course, a lot of, clearly the students are using their brains and doing their own work.

[00:31:58] So that's, for teachers who are anxious about it, I think that is one approach to encourage students. But I'm excited about, I think this is what you were trying to get at, I'm excited about the ways in which it might be helpful to students. And so, I can give you a couple examples of things I’m seeing.

[00:32:18] So, one is, a lot of students, a lot of people have anxiety about writing, right? They find it very stressful. A lot of our students, I teach at a state school, the University of Illinois, you know, something like half of our incoming class, more than half are the first in their family to go to college.

[00:32:36] So, they don't necessarily have models in their family for how do you write a business cover letter? How do you write even a five page academic English paper, right? Or how do you write a science lab report? This may be very new, and there isn't a parent at home who can help them.

[00:32:55] So what the AI can do is, they can give the AI the prompt, the AI can shoot out something that looks vaguely like a lab report, whatever, and then they could use that as a template, right? They could, and I use templates all the time when I'm doing graphics work. Like if I were going to make a flyer, I would go into Canva and like choose from their templates, right?

[00:33:25] And then I would modify it. I would tweak it and tweak it and tweak it. But it would save me a tremendous amount of time and I would get a better result than if I tried to do it from scratch. I don't have those artistic skills.

Charlie Jane: [00:33:37] That’s how I wrote cover letters.

Mary Anne: [00:33:38] Yeah. 

Charlie Jane: [00:33:38] Yeah, for years, that's how I wrote cover letters.

Annalee: [00:33:41] Yeah. 

Mary Anne: [00:33:41] So, there's nothing, I mean, when I first wrote a letter of recommendation, I went and looked for templates. When I write my CV, I go to the department and I ask for a copy of, here are some templates, right? When I'm doing syllabi. Here's, like, everyone who's taught the course before me, we have a file with them that you can reference and look at. So you're not starting from scratch. So, I think that is the big plus is for people who have anxiety or just are at sea, this is a completely new foreign thing to them. You can use, and I say this not having actually done it myself yet, but in theory, you can use ChatGPT to produce something.

[00:34:21] Now, it's interesting. I was in a department meeting yesterday and we were discussing this and one of my colleagues said what’s happening is they're getting emails that are like, the student wants to say, basically, can I get an extension? I'm sorry I missed this. You know, can I get a two-day extension?

[00:34:38] And what ChatGPT will generate is a three-paragraph formal business version of that, right? And so, the faculty are suddenly getting these letters that are like, in very formal business language to basically ask for a two-day extension. So, it’s really obvious.

Charlie Jane: [00:34:57] Since the dawn of time, students have…

Annalee: [00:35:03] Society has always viewed lateness with a negative eye, but actually…

Mary Anne: [00:35:07] Yeah, no, it's more like things like, I really appreciate your willingness to be flexible about this sort of issue and I assure you that I will not do this again. That, like, I will, whatever, whatever, a whole thing. So that's just kind of funny. But, again, that makes them more willing to write to their teacher and ask for an extension, hooray, right? Because that is one of the first barriers. My daughter is 16, in high school, has every academic advantage and nonetheless, when she doesn't hand in an assignment on time, and we say, well, drop your teacher a note. The amount of resistance she has to that is incredible. Right? And we're all like, well, we're all teachers, just drop them a note.

[00:36:03] And, among other things, another piece of this is that our students are really not used to email, right? They don't use it. And so, when. You know, we grew up with it. I mean, I was using it in college or right after college, right? And used it extensively for decades, whereas my daughter, my students, they text, right? They use Instagram. They use various other things they do. They generally are not writing each other emails, right? If they can use ChatGPT for that kind of thing, that’s great. 

[00:36:39] If they can use it, I think, to get over the initial, I don't know what to write my paper about. And hopefully my students were doing it in the classroom, were brainstorming. They have plenty of opportunities. But other classes may not have that kind of scaffolding structure. It may just be, you read everything you get lectured to and then suddenly you have to write a paper on your own at home, right? And a lot of students find that really frightening.

Charlie Jane: [00:37:06] Yeah, definitely I’ve been hearing for decades now, from like, my parents were both professors and I used to hear all the time about that. Every year, they felt like their students were arriving in college less prepared to do that kind of work because high school was no longer preparing them the way it used to. And it's a complicated problem. I was going to just pivot slightly to ask about creative writing.

[00:37:25] Okay. So, when I was starting out and I think Mary Anne, you actually read some of my early fiction and I deeply apologize for that. Like, I think I submitted it to various publications and anthologies that you were involved in like over 20 years ago and my fiction writing was really bad when I started out like I was I was writing really kind of… my characters were a little lifeless. The scenes were kind of dull and I really bought into this idea that you have to write a million words of fiction before you can get good at it. That was the thing that everybody said on the internet, like in the early 2000s, a million words, which sounds like a lot, but actually it's basically like, you can write a million words in a few years if you really crank.

[00:38:08] And, you know, I do feel like when I've written a fight scene, like I've written a dozen fight scenes before, I'm better at writing fight scenes. When I've written a description of a place, I'm getting better at writing descriptions of places. And I worry a little bit that if people are like, it's hard to write a fight scene, it's hard to write a description of a castle, I'm just going to ask ChatGPT to do that part for me and just drop it in. You're not going to get that… It feels very purist and work ethic to me now to say, you have to do your million words. It is very purist and work ethic, but I also think there's some truth to it. How do you respond to people like me who are like, those kids, they should write their castle scenes and their fight scenes. Write your fight scenes, kids.

Mary Anne: [00:38:49] So, I guess I have a complicated response, right? I did classical piano for many years and we drilled and we did scales and a machine can play the same Bach piece, right? And you don't go to the concert to hear the machine. You go to the concert to hear the human performing. So, as long as we care about that, if drilling, right, essentially with scales or running the piece over and over again is going to make the human better able to create this artistic thing, then I do think there's a value in drilling. I'm with you. I'm a New Englander. I am all about the Puritan work ethic.

[00:39:35] So, I absolutely think there's value in that. And when we're teaching composition, freshman comp classes are all about practice, practice, practice, practice, because we're trying to address the same anxiety issue from a different angle. If we just make them do it over and over again, then they'll get used to it and they’ll be able to, they won't be stopped by the anxiety, right? They'll have to do exercises in the classroom, even like the creative writing, 10 minute timed writings at the beginning of the workshop, right, is kind of designed to help you break through that anxiety.

[00:40:07] So we have other tools than ChatGPT that we've developed for helping people get through that. And there's, yes, if you use ChatGPT all the time, if you rely on it, there probably are skill sets that you won't develop, okay. 

[00:40:28] Now, that said, does every writer need to be able to write a really good fight scene? I don't know. You know, like, if most of your book is a Becky Chambers style conversational. I mean, I just read A Half-Built Garden by Ruthanna Emrys, which I love, which is all poly family, people meeting aliens and having conversations. The whole book is conversations, and trying to get through first contact in a reasonable way while taking care of the planet.

[00:41:06] And, yeah, there is like a little bit of a fight scene at one or two points, but it's just not a major part of the book. It's like a page out of 300 pages. Right? So, I don't know. If you use ChatGPT for that little bit, is that different from the artist who starts with a reference image and then, but then puts their own stamp on it.

Annalee: [00:41:35] Or samples something.

Mary Anne: [00:41:38] Yeah, it gets complicated in terms of copyright and plagiarism. And that's almost like another conversation, right? Like, at what point have you transformed it enough that you can call it your own?

[00:41:48] And this is where some professors are being very hard line. Like, if you start with ChatGPT, they are going to consider it plagiarism and this is why I feel like it's a very muddy area right now. And I was telling my students, you have to know what your professor's take on it is right now if you're going to use this, because you don't want to get into a situation where you thought you were just using a tool to help you brainstorm and they think you're plagiarizing. Right? So that’s an issue. 

The other issue, I just wanna be sure to point out, because so many of my students don't know it, is that ChatGPT gets things completely wrong, factually wrong.

Charlie Jane: [00:42:26] Yeah, I was gonna say, and… yeah.

Mary Anne: [00:42:31] I’m seeing exercises that people are doing in the classroom where they're using it to teach critical thinking, which I love. Where they give the students an assignment. They have them use ChatGPT, and then they have them fact check it and figure out which things are verified, which things are accurate, which things aren’t, and then even out of the things that are accurate, which things are actually going to support my argument and which things are not really relevant to my argument.

[00:42:57] So, in that sense—

Annalee: [00:42:59] That’s great.

Mary Anne: [00:43:00] I know, I love it. I think you'll see more of this as some of the more, I don’t know, fast thinking… Fast thinking is not the right word, but the ones who are willing to flex, I guess, and to explore these possibilities and I want to acknowledge the economic issues here, right?

[00:43:23] If you are adjuncting at three different schools and teaching seven classes, and you are already completely overloaded, this feels like a tremendous burden, right? Now it's going to be 10 times harder to keep your students from cheating and you can't redesign all of your courses on a dime, especially coming out of the pandemic where they all had to learn how to teach online and then come back again. 

[00:43:47] So, teachers are exhausted. The timing of this is not great but it's here and I think if we can figure out ways to use it productively, that will enhance our goals of teaching critical thinking and writing skills, I mean, that’s the holy grail, right?

[00:44:07] Like, whatever tools you're using, if you can use them in that way, then you're achieving what you're aiming for, right?

Charlie Jane: [00:44:14] So, I just want to come back to the fight scene thing because I feel like that was maybe a bad example, but I actually think… I'm going to argue that Ruthanna Emrys, I don't know anything about her or her practice, but I'm going to argue that if you want to write a good scene, full stop, writing a lot of different types of scenes is good, and if I have previously written a few fight scenes that turned out well, that's going to make me better at writing a scene where people drink tea and talk about gardening, because I'll just have a better sense of, like, what goes into a scene, how to make it feel vivid, how to make the people feel like they're fully present.

[00:44:48] What I'm getting through what you're saying is that if ChatGPT is being used in a way that encourages people to use their minds and to be fully engaged, to be fully present, we should encourage that. But if it's basically allowing people to just take shortcuts and do what they used to do with CliffsNotes or those other kinds of like, you could look online for sample essays about any topic like 10 years ago, then not so good.

[00:45:12] So, it's really about, like, making sure people are engaged. 

Mary Anne: [00:45:15] I think I would agree, right? I mean, and again, I don't love it when my students use CliffsNotes, because if they do nothing else, I want them to read the actual story. Like, I actually care more about that. If they read a summary of a story, there’s very little for them to engage with, right? Whereas if they read the actual story. I mean, even if it's a terrible story, you know, if you're reading Conan the barbarian, “Queen of the Black Coast” which is full of racism and sexism, but it pops out at you when you read the original text in a way that you might not get from the summary, right? 

[00:45:57] And so if you want them to be able to do analytical thinking in the classroom when they're discussing it. That's the key, is they have the access to the real source material. So, I actually am probably much more troubled by CliffsNotes use than ChatGPT use, in fact. Because CliffsNotes is like cutting them off at the reading stage before they even get to the writing stage, right? If that makes sense.

Annalee: [00:46:20] Yeah. Although, interestingly, one of the things that ChatGPT is being recommended for is summarizing. That's like one of its superpowers is that it can summarize things. So, it could become the new CliffsNotes. That’s really interesting.

Mary Anne: [00:46:35] That's one of the things I would want to warn my students about. Like, sure, you can use this first summary and maybe that's useful. I don't know. I'm not a science person, right? So, like, maybe that's useful for a biology essay or something, I don't know. 

[00:46:48] Although, in fact, when we try to teach them, like, if there's an abstract, you look at the abstract, or if there isn't, there's a summary already there, and if there isn't, look at the introduction and look at the last line of the introduction. That's probably going to be your thesis statement.

[00:47:04] There’s a summary already there. We do build summaries into the essays. But for literature, for art, the details matter, and so, summaries are just not so useful to them, I guess. And it's a process convincing them, especially when they're harried and they're working two jobs and they're trying to squeeze in enough classes to keep their financial aid. Whatever else they have. Elder care and child care responsibility. My students are scrambling and they're working so hard and I have a lot of sympathy for them. 

[00:47:41] So, it’s just a question of how can we best help them get to where we want them to be? To have these skills, to have this knowledge. We can't be afraid of the tools. We have to figure out, like, what are the useful things and what are the problematic things, right? And then, ideally, convince the students so that it's a collaborative learning process. Get them on board.

Annalee: [00:48:10] That's great. Thank you so much for joining us, Mary Anne. That was so interesting. Where can people find your work online? 

Mary Anne: [00:48:15] MaryAnneMohanraj.com is my website. I mostly live on Facebook, although I'm trying to spend more time on BlueSky now. We'll see. And I think I may have abandoned the Twitter at this point. So, yeah. Facebook, if you want to come chat with me, is probably your best bet.

Annalee: [00:48:36] Awesome. Well, thanks so much. And check out Mary Anne's cancer memoir, Tornado, which just came out.

Mary Anne: [00:48:44] Thanks, guys!

Annalee: [00:48:45] Awesome. 

Charlie Jane: [00:48:45] So great to see you, Mary Anne. Bye. 

Mary Anne: [00:48:48] Bye!

Annalee: [00:48:50] You have been listening to Our Opinions Are Correct, and thank you so much for being here. Remember that you can always find us on Mastodon at Wandering.shop and on Patreon at patreon.com/ouropinionsarecorrect. And we're also on Instagram. 

[00:49:06] Thank you so much to our amazing producer, Veronica Simonetti, who makes every episode sound much, much better than it started out as. And thanks to Chris Palmer and Katya Lopez Nichols for the music. And we will talk to you later. If you're a patron, we'll see you on Discord.

[00:49:24] Bye! Did you want to say bye, Charlie?

Charlie Jane: [00:49:29] Oh, sorry. Bye!

[00:49:31] [OOAC theme plays. Science fictiony synth noises over an energetic, jazzy drum line.]