 Let me welcome everybody welcome to the future transform. I'm delighted to see so many of you here today We have a fantastic guest on a topic that I know is a great deal of interest I'm really looking forward to a conversation now our topic today is Teaching with AI and for the past year. We've been very focused on this We've had a whole series of sessions today's session takes an unusual approach And it's one that I haven't actually seen much of in the world and which is one of the reasons I wanted to host this Question is not just how you can teach with AI, but how you can do so creatively So not in a in a regurgitative way not in a rote way But how can we use artificial intelligence tools to expand and enhance our students creativity? Now in order to have this creative conversation happen I am absolutely delighted to welcome an old friend Harry Brown Harry is a professor at DePaul University in Indiana Harry is a wonderful scholar a wonderful thinker a researcher a terrific teacher I've been a longtime fan of one of his books, which is one of his one of the first scholarly books on gaming It's actually about gaming and liberal education and I strongly recommend that but he's done some recent work on using chat GPT and also an image-creating program Dolly To unleash a student's creativity and let me just bring him up on stage Keep in mind by the way. He has just returned from Ireland So at any at any moment not only may he start quaffing Guinness, but he may just pass out on the stage So we got to keep him going Let's see how they further ado. Let me welcome professor Harry Brown Hello, sir Hi Brian. Thanks. Oh So good to see you so good to see you. Where are you today? Are you at home? I know I'm in my office here at DePaul. I just just got in last night. So It was good and I'll try to break the habit of drinking Guinness for the purpose It'll just be an act of cultural understanding sure So I have so many questions for you But but to begin on the forum we asked people to introduce themselves by talking about what they hope to be accomplishing in the next year Now I'm curious. What's on what's on deck for you for the summer and fall and into the spring of 2024? What projects what ideas are top of mind for you? Well, I'll keep it I'll keep it at teaching for the moment since that's the topic of the forum But I'm scheduled to teach a first-year seminar here this starting August this semester and It's a it's a seminar that I've taught before called introduction to luteology. This is introduction to the study of games and play and It's I've taught it before. It's it grew in part from my previous interest in games There's a literary focus. We talked a lot about the Relationship between games and stories, but it's the first time in a few years I've taught it and of course the first time I've taught it since the the advent of chat GPT and I'm adding a whole new module and this is this forum actually I mean Brian and I actually Started talking about doing this form from an email thread on the use of chat GPT in Role-playing games and simulations and so I want to use AI Potentially in having students work with game design and gameplay Specifically in role-playing simulations, so I'm working on that this summer and then I also want to develop a course on Post-truth humor or post-truth comedy and there's there's of course an AI dimension in there and you know using and I'll talk a little bit about a AI and absurdism today But I'd also like to think have students think about the the comic potential for AI. It's Unintentionally or intentionally and if anybody's seen the the recent appearance of AI generated TV commercials like pepperoni hugger fritter synthetic summer you'll kind of I mean, but I kind of want to see the Opening of new frontiers and comedy with AI too. So those are two ideas AI and gaming game development and role-play and AI in Oh Fantastic great topics great topics Friends if you're new to the forum what I'm about to do is absolutely shameless I'm going to seize the mic and ask our guest a couple of questions This is to get the ball rolling and so I'd like you to do as as as Harry and I speak is to think about First of all what interest you have in teaching creatively with AI that you brought to this But also think about what questions you may have in response to the projects He describes and by the way before we go further on the bottom left of the screen There should be a link to a couple of applications one of them is to dolly too Which is the open AI image-generating program that professor Brown used and also of course linked to chat GPT So you may be inspired as we talked to try these out So just just to begin You talked about So many ways of using this but one of them was in a class. I believe on horror and surrealism. Is that right? Right? So what did what did you do? What was your experiment here? Well, I can I can show you I've so I've created some slides And I know the purpose of this forum is dialogue, but I did just want to show everyone please the prompts for the assignments I used to Use AI and and also the outcomes just a few outcomes So I'll just I'll show you the results of the experiment When when we go to the slides, but there were two there were two main things I did with it in that class And it was global horror and surrealism. I had a unit on Japanese folklore and Jay horror and so we talked about the continuity between traditional Japanese Painting yokai stories Contemporary literature and and Jay horror and so one thing that I did is that We started with Japanese folk stories alongside some Contemporary Japanese literature and I had the students kind of like take passages from the the stories and do visual Interpretations using Dolly to and I'll show you some of that too. So think about to think about the way that a particular Archetype in in Japanese folklore works its way through different styles from painting to To to Jay horror and to think about how an idea can be interpreted in in two or three different visual styles And this was key to a couple films and readings what you're doing and the other thing that I did This was this was an assignment later in the in the semester at the end of the semester was I Was kind of bothered by two assumptions that a lot of my colleagues had when chat GPT dropped And which was was mostly reactionary that that students are just this is gonna this is gonna destroy the way that we assign papers and read papers and the humanities and I think one of the problematic assumptions is that Students are natural born cheaters and that the only the only possible thing that they can do with this is Take a shortcut to writing papers and they're not really interested in and learning anything and obviously that's I think that's an ethical problem With with the way that we see students the second one is pedagogical and The assumption is that the the best or only possible response to chat GPT is to somehow AI proof our assignment so that Particular kinds of prompts that can't be done can't be completed using using chat GPT Which automatically assumes that that AI tools? Don't really have a place in the writing or research process except to a short circuit it short cut it Or circumvent it so I wanted to create an assignment That used basically that incorporated I AI as a research method in literature So I called it dialogue with a machine and we did it in the semester so that the students already had a portfolio of of writing assignments that they could cross-reference using Take ideas from their writing portfolio and dialogue with chat GPT about them but also have a working knowledge of different traditions and harm surrealism with which they could That they could bring into their dialogue with chat GPT. So the purpose was to a Take something take things that they had already learned from over the course of the semester dialogue with chat GPT about it so then form dialogue And then see what they could take from that dialogue that they didn't know based on And what we did know and then to propose Possible research paths based on that dialogue So this is this is of course more constructive So using using chat GPT to give them ideas for future research rather than just using it as a paper writing substitute Nice very for a constructivist. Yeah So do you want to put up one of these prompts or sure? Yeah, and there's a brian I did a right before about 15 minutes before I sent you the link. So this is uh, there's There's a few slides here and I'm gonna I'm not gonna spend more than a couple seconds on on on each one So I brian will will post the link and mainly I want you to use these as reference points for the for the questions in the form itself But I felt like I had to show you the prompts and the outcomes so that we could have a Kind of firm basis for the for the dialogue. So yes So you got the you have the link brian, right? I have a link, but it wants me to it wants you to approve Be using the link Um, oh hang on. Hang on. Uh, I think this now works Let me just put this quickly in the chat Yeah, I think google duck link and also, um But friends you should be able to see this now on the on the in the chat box Um, and already that the chat is flowing with conversations and questions about this, but um, good Good, so the slide I'll go I'll go very quickly. I just want to show you what I did so that we can We can talk about it concretely. So all right. I'll share this uh, and I'll um There we go You might want to do the slideshare. I'm gonna do. Yeah. Yeah. Yeah. So, okay there. Okay. You got it Uh, yes experiment one experiment one. Okay. So this was an intro to lit class this semester the the topic this is uh over here is the uh Is a slide that I shared with the class and this is just jabber walkie in the uh The lesson the basic lesson was the relationship between uh form and meaning in poetry and uh the way that form itself Can create or layer meaning on content. So this is uh, this is uh, of course, uh, uh a nonsense poem that gains its sense because carol writes it in traditional ballad form. So we're talking about ballad this this day and uh So how can form layer meaning on anomic content content that is uh devoid or confusing in its meeting So the chat gpt prompt and we goofed around with this in class was to write a ballad ballad form And this was february. This is early february. So it was the day that the that everybody was talking about the the chinese spy balloons And uh simultaneously with oh right taylor swift breaking the internet So this is this was just how do you how do you take anything any content? impose a form on it And and how does the the meaning of that content change? through the imposition of form so um Ballad style taylor swift Chinese spy balloon So really we just use chat gpt to marry the news story of the spy balloon with uh taylor swift songwriting technique And uh, this is this is what it gave and again, I won't read all this you can browse it The the pink highlights are things that we talked about in class that were Lines that that had a certain tailoredness to them. Uh, and so we we kind of like talked about The ai's ability to read and uh taylor swift like what was taylor swift so form and meaning this was uh A different form uh anglo-sax and a literative verse we we've done baowulf and anglo-sax and riddles prior to this so Yeah, and this is uh the highlights here are the uh the alliteration and uh, it even incorporated the caesuras the uh The line in the middle of these so you see the so the two main things we talked about with a liter of verse Whereas the alliteration and uh and the caesuras and the balance between the kind of like two halves of the Of the alliterative line and so this is what it did And again pretty good job, but again the the the point was that uh form and meaning are both interdependent and independent of each other That you could have To uh to different kinds of content so experiment to this was later in the hang on Hang on question. Just just a quick practical question before you get on that Did you do this as one instance of chat gpt on the projection for the entire class or do each student get to uh Try this out. This was my demo, but students students also tried it out themselves So we had in fact, it was a student who suggested that we like I said just what what what what do we want it to do And it was a student who suggested the spy balloon thing because they were talking about class so it was yeah And uh same class. This was on dramatic structure Uh, and we were doing hamlet with rosencrantz and gildenstern are dead So certain some exquisite corpse This is the basis was just a class slide basic dramatic structure of hamlet. Uh, how does uh Tom stoppard play with this Structure especially considering that one of the predominant refrains and themes in the play is not knowing rosencrantz and gildenstern do not know The structure of the drama participating in so we talked about exquisite corpse and uh, what would it look like if you applied exquisite corpse to Traditional dramatic structure. So we asked chat gp to generate 18 inciting incidents for potential short stories or short plays involved tragedy or suspense and then what we did was uh, so this was It gave us this And what they had to do is pass these sheets around in class physically You know how exquisite corpse works So that they can only see the uh, uh, this the stage of the drama that occurred immediately before but but not before that so exquisite corpse with uh dramatic structure And this was one of the results and you can see I wanted to take the picture of it and show you how they're folding it and passing it but uh, this is It doesn't make sense because it's not supposed to make sense You're supposed to write a narrative Not knowing what we're internally the different parts of the narrative don't know what the other parts are doing Because this is what stopper does in the play And you can see how this all starts from a cat sneaking into a bedroom And it ends with uh waking up from a coma 13 years later with no memory It takes a really interesting series of turns just because the students just like rosencrantz and gilden stern don't know what's happening Uh, it except what they're seeing right in front of them. They don't know the structure They don't they don't know the plot and so what you get is a plot That is an exquisite corpse, uh That that kind of goes nowhere and somewhere at the same time. This was another one Uh, and this one was interesting. You see here the the this is like Characteristic of what happens in stopper's play because here Uh in in one stage, uh, you have intentionally the speaker trying to shoot his father And then because they uh are here up here in the complication Uh, and then in the additional complication. He's trying to save his father So it is absurdist because he shoots his father and tries to save him at the same time Uh, which is kind of uh, there are moments in stopper's play in which those kind of mixed intentions happen And of course, I love the wrap-up where the doctors replace his brain with a cd of shrek 2 So that means we live a normal life We just had the students love these things. Uh, and as I said, they're supposed to be nonsensical They had a great laugh And then the next stage could be if it were a creative writing class to take one of these exquisite corpses And to shape it into an absurdist drama or an absurdist story in the style of tom stopper. So This was for a literature and ecology class Um speculative this the the the topic was speculative ecology and again, this was the prompt We were reading peter ward's life as we do not know it which is Ward is a Uh an astrobiologist who speculates on the potential for life Uh different places in the solar system So we'd read that uh, and then we were we were storyboarding a hypothetical series Uh, uh Documentary series exploring the potential for life elsewhere, uh in uh in the solar system So this is the the role play is that they're working for netflix to develop a series and they're storyboarding some concept art using dolly 2 So that they can pitch this, uh, which the cgi team will will use to create background art and animations for the series So this is uh, this is the prompt right here We did this, uh, we did this in class one day and then kind of like Discussed it talked about it step five On uh on the next day and of course it was dolly 2 But it's informed by our reading in life as we do not know it and uh, these are some of the cool results The keywords come from ward's book Deep ocean vent fauna Europa jupiter. These are all uh dolly generated images. You can Be more specific. So uh different exotic life forms that ward talks about like syphonophores We added in we know we refined it so that we could Just have a clear concept of what these Exo biological creatures might look like and again, you can storyboard it using that And this is this is what i've already talked about japanese folklore, uh j. Har This is the prompt and uh basically Take a consistent theme and reinterpret it in different visual styles and then talk about a traditional japanese painting Japanese horror films these are the three stories we were talking about Um, and uh, this was these were some of the results. So an abandoned village. We're working with uh, hakiwara as the town of cats Uh an abandoned village filled with feral cats and the style of traditional japanese painting And then you translate that same idea into, uh style Japanese horror film Yeah, so it's kind of it it's and we talked about like what are the j. Har aspects of this of these images And how do we how do we see those images? How do we see those? Styles kind of mapping into films that we watched in class Like like the ring and and this this was my favorite actually so it uh, it was responding to the same prompt visual interpretation of Of hakiwara's story Except instead of showing uh populated by feral cats it showed their presence indirectly by uh, lots of dead rats It was I thought I thought it was a weird Leap and the students loved it too where it's like they're not showing the cats But they're showing the results of the cats Indirectly the results of the cats presence with these scattered the bodies of rats So we talked about how this was like this would be great You know a great technique to use if you were kind of like making a short film based on the town of cats Uh as some kind of an uh, if you wanted And on it, um, so again different visual interpretations of of litter and I didn't I don't have it right here But the students had to isolate A series of quotations or a quotation from the story that you're using to prompt the ai To uh to interpret a particular moment in the story and this is the last one Um This was a ai research method. So this was again in in part a response to my own colleagues who said that uh um that chat gpt will Um, that's going to destroy the way that we work in in literature and philosophy classes And this was uh, so again, you can read this on your own on the slides, but this this was the prompt Uh, basically use chat gpt to uh bounce off of what we learned and tell us Something arrived at some insights that we didn't learn based on what we did through this dialogue So the paper consisted of uh A dial the students had to kind of like edit their dialogue Taking excerpts from uh what chat gp told them and of course the creative part of this was that the students had to Prompt them ask good questions in order to get good results and this was our class demo We generated these questions together And uh, it starts with a very general thing The the basic topic of the class we kind of ran through I'll just I'll go through these I wanted you to I just wanted to record on the slides how it all played out but we kind of moved from uh Definitions specific definitions by Freud and legati on the uncanny to a kind of broader definition of humanness and what humanness is Uh to a discussion of uh of ai and humanness consciousness the nature of subjectivity is really interesting and then of course the uh The outcome the the purpose of the assignment was to uh Uh again create future directions for research and based on those slides, you know that I just kind of ran through there There were three as a class we kind of like just from the demo we extracted Maybe three topics that I could add to this course, uh, or write about the the next time I approach it one was Uh disability in in horror and surrealism, which we didn't talk about much over the course of the semester genetic mutation Diversity and ai tools are making so these are these are ideas that were not included in the course at all Of course, they're they're interesting And it's the dialogue that kind of like led us to say yeah We could take what we did here and then maybe the next time we talk about this Uh, we can incorporate these topics So that was that was the point to use it constructively rather than just as a as a covert tool for a Completing tax. So that's it. I'll on share but uh, wow. Yeah, thanks for very well Well, that's fantastic. Oh just People are going to have an avalanche of questions for you, but I want to ask one quick one myself, which is um Was this something that you did entirely on your own? You know grabbing dolly grabbing chat you bt Introducing the classroom or did you need support from either your department or from it or anybody else? No, I did it on my own Of course, you know the the great thing about open ai is that it's it's just there and the students could access it instantly in class It was just yeah, and I don't think I could have done it You know, it's I mean it It made it easier that we could just kind of like Just jump into it in class. Uh, so no, I didn't have of course It was it was supported in form with lots and lots of conversations between fall of last year and spring of this year On uh on on chat gpt specifically, but uh, no no formal no formal support I just kind of students were always talking about it too. So They were just kind of like naturally motivated and intrigued by uh by the work Oh Very nice. Oh, this is this is terrific All right, I'm so excited to see this but they let let me get uh get out of the way and And provide opportunity for people to ask questions Friends if you're new to the forum remember again the very bottom of the screen Is that white band with a few different options and the click the raised hand if you want to join us on stage And otherwise click the question mark to type in a question In fact, we have a couple of those right now. Uh, so let me just splash one on the screen Hang on a second And this is one from uh, our dear friend tom hams who has a habit of asking very elegant and deep questions Tom says generative ai seems to demand a higher level of thinking on blooms texanui The many of our students and workers are trained to do How do we need change teaching to raise their level of imagination? huh Well, I mean one thing that was that was useful for me I didn't start with this broad question of how do I use ai in my teaching because we're all going to have to reconcile with it You know one way or the other. I always subordinated Um What I was doing with dali or or chat gbt to the the content and purpose of the of the course itself and the specific day So it was always I mean you can see in the in the slides. I kind of like put the um, you know the Learning goal of the day first and then you know And then like how how do we use ai to serve that particular? Intellectual purpose other than making ai itself the uh the the intellectual purpose and then kind of subordinating so all my examples are disciplinary uh to uh to literature And uh every every course would would you know would have its own purpose. So I would say thinking Thinking would begin first with uh, whatever the the purpose of the course is and then ai Serves that purpose as a tool rather than becomes a kind of like dominating Driver of of what you're doing. So We did talk about You know in in the classes with the students the way that That these tools could be an enhancement to imagination Imagining a certain thing and and one thing was a particular enhancement like okay, we can't go to uh We can't travel beneath the ice crust of Europa Not yet to uh to uh to see what kind of uh Organisms might be there or we can't go back Four billion years into the past and see what mars might have looked like when there was water there. So We could we could take ideas and and give them a kind of an imaginative structure using using a Dolly specifically so I think they can act as enhancements to imagination As long as uh again for me at least uh, they were they were subordinate to The particular goals of the course on that particular day and subordinate to the readings that we were doing that day So we didn't do any of these things That uh that were not connected in somehow to the reading assignments for that day And so it didn't exactly change the way that I constructed a course like a literature course It was how do I uh, how do I teach the stuff that I would have included anyway in a new way? That's that's a very very thoughtful answer. Thank you Thank you. Uh and top mess always. Uh, that's an excellent question Friends, that's an example of a q&a now. I'm going to try and uh and bring in A wonderful scholar from new york Clay shirky to ask a question of his and do this on video. So let me see if we can get everyone together and Together. Yes, hello clay. Hey, good to see you brian. Thanks so much for this. This is super interesting Um, did you you talked about the students being really engaged with these kinds of experiments? Did you as you structured the assignments asked them for any kind of after the fact reflections on what they'd learned from the experience How they might use these tools in the future any of that kind of You know after the fact cognition about the Assignment itself. Yeah, that was that was actually a big part of it. You can see that uh, You know that if you look at like number five on the prompt on speculative ecologies It was it was a kind of like meta reflection on Um, you know, what could what did we learn from this? What are the limitations of it and we actually spent a lot of time a good week in uh At the end of this past semester in the in the hard surrealism class on talking about ai as a research tool and What what they gained we demoed this in class we had it was it was a writing intensive a w class and we uh We workshopped together With chat gpt in class over a series of days and kind of like talked about how to craft a good question What does a good dialogue with an ai actually mean? How do you miss it productive responses? What kinds of knowledge could you bring into it all that stuff? So yeah, it was I mean, I think we're just The students had the sense and that we were not making it up as we went along But then we have to rethink the way that we do research in a dialogic way Using a tool like this other than just kind of like typing in Like, you know research in the humanities didn't end when we went from, you know Physical libraries to search engines and so it's this is I mean I kind of stress the dialogue thing It's not going to work. It's not going to write It won't work unless you bring some good prompts and good thoughts into it and knowledge that you have to have Good knowledge in order to ask good questions. And so we kind of emphasized this and uh, like it wasn't Like they they came to realize eventually that they it wasn't for like outsourcing their responsibility for writing a paper But they had to bring some responsibility An awareness to the assignment Because it wasn't going to work if they if they couldn't ask a good question So kind of get back to the ancient thing like what is a good question? And we talked about that too Great, thank you Oh, thank you clay. Thank you very much. And please stay safe and uh in a cp atone new york Um, thank you We have uh, so friends if you're new to the forum So you just saw an example of a video question And if you want to follow clay just hit the raise 10 button and i'll put you on the list And then you saw tom hams's q&a box. So please both the q&a box and the raised hand are there for you As we go. Uh, so please share your questions and especially your your practical ones and your theoretical ones and Um, and how are you can even ask what are some other ways to show the presence of cats invisibly? Um While you're thinking about that, uh, we have a question from uh, dear carl haca renan coming to us from new england Uh, and carl asks each spin of the chat produces a variant. How do you deal with the challenge of non reproducibility? Yeah, yeah, that's a good question too. So when we um, uh, We we started with this kind of like broad concept of the uncanny which which ran through a lot of discussions in our course and uh, it was a kind of Non reproducibility. So it gives often similar answers, but but not the same And uh, if you have say 20 students in the class asking Exactly the same question but receiving 20 different variants Uh, uh in response One thing that we could do is kind of like compare the responses synthesize and extract, uh, the best ideas or consistent ideas that that came from, uh Variants to the response and we did that like you could just say what are your thoughts on the uncanny Or and it's using film or something like that Different things and then we could then we each student could, uh Kind of like share the response that they got we could look for patterns in the responses talk about the particular patterns that emerge Talk about the variations that emerge One of the things we did in the workshop is that I asked the student to share Like the most interesting thing that came out of their series of questions. So The non reproducibility for us was uh, was was just interesting because it allowed us to see both pattern and variation And to kind of like learn from both and emphasize both one of the strategies that uh, I suggested in the course of a longer dialogue thread was to uh, Identify the patterns and variations and drill on those things like keep asking about like Human the human what is it? And that's that kind of like That led us again down the more specific we got the more interesting it got I guess Uh, that's a good takeaway. That's a very good takeaway Uh in the chat, uh, rachel barlow is um, is uh, has instigated a discussion about the prompt engineering Which is very interesting. Uh, and we take a look at that afterwards, but um, But one of the things she points out is is is trying to figure out ways either describing prompt engineering or doing it Well, and uh, greater specificity seems to be a wonderful takeaway. I just saw an interesting experiment where someone was using chat gpt To produce detailed instructions for mid-journey and the problem is that mid-journey appeared after the chat gpt dataset ended So the prompter wrote a two-page description of mid-journey Uploaded that so taught chat gpt what mid-journey is and then had mid-journey provide incredibly detailed Down to coding level prompts for mid-journey, uh, which worked surprisingly well This all sounds a bit that would have been an insane sentence to say about a year ago But now this just seems kind of kind of prosaic Um, we have more questions coming in and carl. Thank you so much. Uh, that's a very very good question Uh, we have another follow-up on this from our friend in madison Who hopefully isn't also in the smoke plume? John Hollenbeck asks what media attributes are uniquely assignable to generative ai? In other words now that we have chat gpt we can Do what? Media attributes. Um, I used it a lot a lot of what I used it for is kind of like Storyboard so maybe trans media is is one thing that I use You take this story and imagine it to imagine it visually couple of the applied assignments I had in the uh, uh in the living systems course, uh, was to Again storyboard a documentary based on Things that we learned about but but couldn't necessarily see I haven't experimented with the music ai yet or the or the video generators yet But I mean that's That's coming if not already here. I mean I worked with again text and uh an image And I did find it useful chat gpt can't do image and and and dali can't do text so It was interesting to kind of go back and forth between the two And uh, I think it would be I think it would be very useful in Any kind of course that involved any kind of trans media component Ideas concepts from from one medium into another medium And that's why it was so useful and in hard surrealism because we were reading a lot of stories But also discussing films, uh, and even discussing the way that uh that we might make films You know based on Our knowledge of horror and surrealism, so yeah, um But uh, I mean you you saw what What can we do with chat gpt now? We can take apart students can easily deconstruct forms poetic forms narrative forms dramatic forms and it was just fun to kind of like see that and I don't think it Like it didn't if I say, you know write a sonnet as if Or you know write a taylor swift song about the chinese spy balloon. I don't think it What it shows is that these forms at least which which could we used the textbook that we used was strand and boland's the making of a poem Which kind of kind of marched. Is it really good? It's really good first year intro level textbook on poetic forms But it is it so it showed that you could kind of like take the forms as strand and boland were describing them And just play with them in a whole lot of different ways. So for us in that class Uh, it kind of showed the the by literature as play Uh, it allowed us to create all these allowed us to use chat gpt to create All kinds of form, you know variations on on these different forms Which was important for that class because you have first year students lots of non majors lots of gen ed students and uh It was a way to kind of uh enliven Something, you know a very traditional approach to poetic form in that class. So Nice nice for non majors in particular and and people just so new to this Oh, that's excellent. Thank you. That's a that's a great answer We we have more questions coming in harry and and I want to cobble one together That comes from our wonderful friend and longtime chat meister Lisa dirf and lisa pointed out That students not only have uneven experience and understanding of literature and surrealism But they also have a very uneven background when it comes to chat gpt and dali And so I guess I'll put that as a question as a teacher How did you handle that? You know when you might have a student who's already written six essays? Using chat gpt versus one who's never fired it up. How how do you handle that that unevenness? Step by step and I you know, I don't normally put my prompts in steps one two three four five But the slides that I shared we did this in class. So these are classroom prompts not necessarily Paper prompts and uh, we did I mean we started from zero like open chat open Dolly to some of them who had more experience said well, I like stable diffusion or I like night What can I do at cafe? I like working with that and I did I just let them go You know, it's so with a lot. It didn't have to be dali We did we kind of like this was a this was an interesting med experiment But we kind of put the same prompt into dali and stable diffusion and we kind of like talk Which ones we liked or why or which one which one was truest? to The reading that we were that we were trying to visualize But no, I did it. I I assume that and in fact some of this when I said Anybody work with uh, you know dali to or You know half them would say no, I don't even know what it is. So I started from zero and Simple answer is just a good old-fashioned, you know spending time in class. So open it Here's how here's where you put in a prompt open an account with open ai And play with it for put the stuff up on screen Walk through some demos put them into groups in class Maybe combining the students who have some working knowledge and or not and uh, just working through it in class. So the Uh, the speculative ecology assignment that was a that was a two-class assignment where class one was um, learn the tool Generate some images class two was discuss what you've generated in relation to towards book. So it's just I just take kind of a workshop approach Oh, that's Thank you. That's really really helpful That's Yeah, this is extremely helpful and for those of you revealing this on youtube some days months years afterwards I think this is this is a really really helpful Guide to using this. Thank you lisa for the prompt We had a follow-up comment from jd in the chat who says there's also the issue of equity Not all students can access chat gpt anytime they want unless they've purchased a subscription There's also concern when it comes to using chat gpt in class I'm one that was a comment not a not a question But I'm wondering if you if you want to speak to that at all Yeah, that's actually one reason that I use three instead of four because They could do it without a subscription and they could do it instantaneously in class And so yeah, I was I was kind of mindful of that I didn't require them to use anything that wasn't available for free You know as reasonably as possible You know as accessible as possible. So yeah, that is That that's that's going to be an issue ongoing because there was a student who's really really into this Who kept saying we should be using for we should be using for so many and And I just I mean I just had to tell them but I don't I don't Yeah, I don't want to require every student in the class to get a subscription In order to do this. So I kind of like stuck to what was accessible available for free Well, that's very smart. That's very smart jd. First of all, thank you for mentioning the the equity question And and you know for what it's worth here. I agree that your student four is way better But but your approach is is the really equitable one We have a follow-up practical question along these lines This comes from our friend Shelby Rosengarten in florida Shelby it's always good to hear you and For once we get to be north of you having more problems than you south of us So the question is these creative exercises are fascinating How much did you talk about the ai and how much did you still consider just discuss the course material? Did the ai take up more time than the original subject matter? No, and I'm not an ai expert And again, this is uh, you know, I learned this Myself just by playing with it. And as I said in response to one of the previous questions Uh, the course material is already is always primary For me And there's there's lots of stuff that I don't I don't apply ai to every single thing that I teach It's just these particular five experiments kind of made sense based on what I was doing Uh, especially in horror and surrealism where we talked about like You know the blurry boundaries of what we consider to be human And other so In that sense ai kind of naturally worked its way into the conversation We talked about the turning test in the uncanny valley all that stuff in that class Intersections between technology and you know traditional problems in in surrealism and um So yeah, it's again I haven't I haven't I have a very limited technical knowledge on ai and I think the great thing about open ai is that again, it uh It breaks down a lot of those barriers so that non experts can do fun things In a variety of disciplines in a class with little preparation or technical support or funding and and you know in service of uh of Of learning a new learning goal. So we didn't uh, I didn't I didn't pretend I said up front Again, I don't have any I don't really know how this works. This is a black box for me And so we did talk about how to use it and how to use it well But we didn't talk about it was always like What can we learn? From this on this particular day in connection to this reading whether it's speculative ecology or japanese Um, you know yokai stories it was always uh the the purpose the content the course content always came first for me And I didn't I only talked about ai in so far as the students had to know how to use it in order to get to What we wanted to learn about x on that day Well, that's that's a really solid answer Thank you and shelby. That was a great question Thank you very much one quick question for me on that harry Did any of your students have a computer science background that they want? Yeah, they did Yeah, that guy who kept telling me that we should be using for uh, it was He's also but he was he was great because we had all these conversations outside class about, uh, you know Not just like ai creating npc dialogue in in video games, but uh, You know, it's kind of improvisational capacity for uh for chat gpt to like lead a dm You know kind of like dnd style like tabletop experience And uh, that's where you came in brine. That's that we we were kind of like talking about that And that's why I responded to the thread that you shared um, but yeah, he was uh the ones that know a lot about how this works, uh, uh are are useful Uh, but again, not everybody not everybody needs that and again I got into this that that guy is actually gonna be kind of an unofficial ta for me in that seminar because he does Now what goes on inside the black box? So he's going to help me develop some assignments on um on Chat gpt in rpg In role-playing games Yeah, yeah, oh, that's that's well first of all, that's a lot of acronyms to to work through I'm impressed But also that's great for him. Um, especially, you know as an undergrad student This is a wonderful liberal arts approach Um, and I'm so happy for this person Um friends, we've got about eight minutes left and I want to make sure everyone gets a chance to ask their question And we have a video question coming from another good friend Previous guest great consultant writer thinker steve airman who's coming to us from maryland um, let's see steve How are you doing today? I'm doing well except for the air quality. Oh man, stay out of it. Stay out of it As well as just it looks misty outside Yeah, same here I was stimulated by the the previous question from shelby um to wonder uh A in the context of let's say a ba in english or In the context of general education Is it as you think forward a few years Um, are these tools important enough as aids to thinking of one sort another? A student's sophistication that using them for that end ought to be developed You know over two or three courses um The other future being Where I I think it is now which is faculty member who thinks it would be good for one course They use it for one course, but there's nothing Coordinated about developing skills. Yeah, I don't advocate one or the other. I really look to you with your experience What would you That's that's that's a good question. Uh, I think as it becomes As it develops, uh, and I don't know if it's going to plateau Whatever that means, but it's uh, I like the idea of treating it as a competency So at the paw for example, we have three competencies writing speaking and listening and, um Quantitative reasoning and so every student Has to have at least one course in one of these competencies and they could you know, any number of courses could complete that competency I've always advocated for a fourth competency in technical literacy or I'm but I'm not like a fourth competency t And I think the advent of of ai almost makes it More more necessary to think because there is no discipline in which Ai can't be somehow applied Even even in a discipline that seems very distant from like like literary studies I think that every discipline in its own way is going to have to think about its own future as with this new thing and in its you know in the kind of Ecosystem of the way that we teach and learn and research. So Yeah, I do think over the long term Everybody no matter what you do in the classroom or in research is going to have to think about this and a dedicated course or a series of competency courses In in in t in general or or ai specifically might be useful If a particular institution Given the little You know Bandwidth in its curriculum. So but I wouldn't be I'd be interested to see how that develops You know over the next like five years Yeah, ai competency, and that's that's kind of an interesting idea Oh, thank you steve. Hi. That's a great question Uh, I appreciate of course the future-looking aspect of this We have a another future oriented question from david furlough in the chat who wants to know Did you get any pushback any resistance to this from you know? I would have for you from students from colleagues from people off the street Students. No, I mean they said Students it was the opposite They're like we wish that more of our courses were like paying attention to this because How can you not so Even I had a lot of music students who were like telling me about like the way that ai was changing The way that the people compose and they get that's completely beyond me But it was so pushback didn't come from the students. They almost I mean if One of my classes were here responding to the last question. They would say yes, we all need ai competency because it's We want to do I mean they didn't they didn't say this straight out, but I sense that they saw more potential in it than Like we just want to use it to cheat, right? We just want to use it to write papers I mean they they see it as something powerful And they they kind of want to develop it colleagues as I said Um Lots of them Just were like, how do we ai there was there lots of kind of anxious email threads Is this does this paper sound like it was written by chat gpt? What are some markers that I should look for potentially in papers that were written by chat gpt? How do I craft an assignment that is ai proof? Lots of stuff to avoid So nobody pushed back specifically on on what I was doing one reason I I love to paw and have you know uh Stayed here for 20 years is because they give us relatively free range as far as course development uh and teaching so it's um I didn't have any uh Any pushback from from colleagues other than You know other than the kind of like starting assumption that that some Still haven't gotten past, uh, which is that this is gonna like obviate our profession or obviate the art of writing Or uh, it's it's gonna create, you know a bunch of drones who just kind of like outsource their thinking to uh To an ai tool. So that may be that may be the kind of like indirect pushback like, uh, How do we instead of asking how do we use ai and what we do or use it to enhance and what we do? How do we ai proof what we do which to me seems a little counterproductive? Well, it's interesting. I mean it's only been a few months But you're kind of seeing a divergence of faculty into very very different schools of thought Um, I think so. Um, yeah, I mean one of the reasons I started doing this Uh, initially was just to try to like think through that assumption that Students are going to treat and uh, we're not going to be able to teach writing anymore because of this. So yeah Yeah, is there is there another answer? to Writing assignments that have them talk about like personal experience Their own personal experience which chat GPT couldn't fake Or, you know developing a rubric for like spotting ai markers Tell students like not to have in their papers. It just uh, yeah, I mean it's um Yeah, like you say, I mean creative teaching. How do we teach creatively with this rather than trying to dodge it? Harry the one thing that we can't do is dodge the end of the hour, which is about a minute away. Um, and I want to seize the podium for for myself for a second to ask you one final question Looking ahead at the next academic year, you know looking bringing 2024 um, I mean we You seem to have articulated one path for instructors to use in teaching with generative ai Uh, this this kind of constructivist approach that's very creative approach one that has an eye on equity One that's very playful in the chat rocks and ricks and rick bands that everybody Think about this more playfully educators should be more playful, which which I agree um Do you What are some of the ways you might imagine this going? Um, so the faculty following your footsteps in other disciplines, you know everything from say economics to biology astrophysics and what kind of constructive playful uses of ai might we expect? music and video uh, and that's uh, it's I mean advertise you can see it's there's you know the the synthetic summer and uh, and pepperoni hug spot There's there's spoofs on ai generated video in order to I mean both of them are pretty deep in the uncanny valley if you uh, if you look at them Yeah, it can't do fingers. It can't do like eyes or mouths or so it always looks weird like those cat Yeah, but I think again as it gets So the student in my class who kind of like knew about a lot about how this work. He's like it's learning It's learning, you know the day by day those faces are going to get better and it's going to kind of it's going to learn itself out of the Valley, so I think uh video composition Is going to be a an interesting way to play with this You won't have to search some clip on youtube, uh in order to uh Make some you know point in class that you want to demonstrate with video Maybe there'll be a version of ai where you could say In class i'm i'm working on the connection between yokei yokei stories and j. Har. Can you generate a three minute short film? Applying a kutiga was the hell screen That we could talk about in class so Stuff like that and music of course I think that that would be cool, too Like maybe you could use some kind of a music ai program to create a soundtrack for a short film that you could As an adaptation of a particular story that we you know that you're using in class So yeah, I think music and video are going to be a big part of the the near future and uh again I do believe in uh, uh, you know finding New uses for it through play And uh, again the more we play the more it learns from us the more we learn from it and uh Again, I don't think I don't think it's going to annihilate humanistic humanistic teaching and learning Well, it seems like it's giving us new venues for humanistic teaching and learning. Um Harry you've been fantastic. It's just a delight to see this project And i'm so grateful to you for sharing all of your thoughts and your practice for this hour Two two questions before you go one is can I share that uh google drive link? Yeah, sure. That's why I sent it. Yeah. Thank you. Thank you and and second How can people keep up with you when they want to find out about your full class? I have a smaller non-existent social media footprint, so uh That's partly by intention. I don't always trust myself on social media, uh, but uh, You can keep up with me by uh by email Very good or uh Yeah, anything. I mean just you you know where to find me Uh, I'm I've been at the paw for a while and I'll be here for the foreseeable future So just reach out if you have any questions observations or potential collaborations. I mean, I'm really open to uh to uh Collaborate on a lot of this stuff. So Well, that's a great call and one of the things we do here in the form is is networking of getting people together for these kind of projects So, um, this would be great and nothing else. I want to bring you back to hear about your next steps in this Yeah, sure Thanks. I'll be happy Thank you so much and good. Good luck with everything here. Make sure that especially that your uh internal clock readjusts to uh In the end of time. Yeah, and and take care Thanks, brian. Thanks everybody happy Yeah, indeed But don't go away yet friends. Let me just point out where things are going and uh, thank you By the way for that torrent of ideas and questions I hope you enjoyed this session and I found a lot of rewarding stuff If you'd like please reach out to harry you can tell he's very very accessible very kind and if you want to talk about this More please hit us up on twitter or mastodon You can see our our logins there and and just use the hashtag f tte if you'd like to keep Pursuing this topic if you want to look back into our previous sessions on among other things teaching and of course AI and of course creativity. Just look at our archive at tiny url.com slash fdf archive We have more sessions coming up everything On college teaching and more on AI and campus economics. Just go to forum.futureofeducation.us to see more Thank you all for a great conversation. I really appreciate all of your ideas Listen, I hope everybody stays safe in this wild summer right now And good luck with all of your summer work. We'll see you next time online. Take care. Bye. Bye