From Pain to Creativity: How AI Helped Kristina Kashtanova Illustrate Her “Zarya of the Dawn” Story – featuring Richmond Law’s Ashley Dobbs and Roger Skalbeck (TGIR Ep. 196)

Marlene Gebauer, Greg Lambert

In this episode, we dive into the fascinating story of Kristina Kashtanova, author of “Zarya of the Dawn,” a comic book that she illustrated using AI-generated images. Kristina shares her personal struggles during the pandemic, including losing loved ones, being unemployed, and undergoing dental surgery. She talks about how she discovered the power of AI-generated images through OpenAI DALL-E and how it helped her overcome her pain and isolation. We learn about her creative process of generating various images before remembering her story and deciding to use AI-generated images to illustrate it. Kristina also shares her experience of sharing her progress on social media and receiving positive feedback from the AI community. Her story is a testament to the intersection of art and technology and how it can be used to overcome personal struggles and create something beautiful.

We are also going by a duo of Richmond Law School Professors, Ashley Dobbs and Roger Skalbeck. Professor Dobbs runs the IP and Transactional Law Clinic at Richmond and explains that the clinic provides an opportunity for law students to work directly with clients on intellectual property matters, such as copyright and trademark protection, under her supervision. The clinic primarily works with startups, entrepreneurs, and creators who cannot afford legal services. Ashley and her team also handle various transactional matters related to intellectual property, such as forming entities, reviewing contracts, and assigning rights. By working with clients in a real-world setting, law students are able to apply their book learning to practical situations and gain valuable experience before entering the workforce. She is also providing assistance to her fellow professor Roger Skalbeck for his “Copyright §101” Comic Book.

Roger Skalbeck created a comic book to teach his students about copyright laws in the United States. Roger explains that he wanted to create something that looked like the comic books he grew up with, such as the Avengers and Spider-Man, with a vibrant and simplified aesthetic. He used various tools, including Mid-Journey for image generation and Photoshop and Pixlr for updates, before putting it together with a layout program called Comic Life 3. Roger’s comic book provides a visual representation of each individual definition in the statute, making it easier for students to understand complex legal concepts. By using a comic book as a teaching tool, Roger is able to engage his students in a fun and creative way while also providing them with a valuable learning experience. Tune in to learn more about Roger’s efforts to use a comic book to teach copyright laws and how it is helping to transform legal education. Roger has a class set up in the Fall Semester this year that will require his students to create a comic book that focuses on a practical aspect of Access to Justice.

Tune in to learn more about the intersection of law, comic books, AI, and copyright. Make sure to subscribe on your favorite podcast platforms and share the podcast with your colleagues.

Listen on mobile platforms:  Apple Podcasts LogoApple Podcasts |  Spotify LogoSpotify
Contact Us:

Twitter: ⁠@gebauerm⁠, or ⁠@glambert⁠
Voicemail: 713-487-7821
Email: geekinreviewpodcast@gmail.com
Music: ⁠Jerry David DeCicca

Transcript

Marlene Gebauer 0:07
Welcome to The Geek in Review. The podcast focused on innovative and creative ideas in the legal industry. I’m Marlene Gebauer.

Greg Lambert 0:15
And I’m Greg Lambert. So Marlene, we are jumping into a topic this week that I know that I love to talk about. And that’s, you know, any, any chance I get a chance to talk about comic books I’m going to talk about if I get a chance to talk about AI, I’m going to talk about it. And we get to blend those together. And really kind of see how each of them can kind of relate and work together in the creative process.

Marlene Gebauer 0:42
Yeah, I’m really looking forward to this one. You know, we even we even got like special prizes, you know, in the mail. Yes, and release in relation to this podcast.

Greg Lambert 0:52
So that’s true. So this week, we have Roger Skalbeck who is professor of law at the University of Richmond Law School. Ashley Dobbs, Professor and Director of IP and business transaction Clinic at the University of Richmond Law School. And Kris Kashtanova, AI consultant, and author of Zarya of the Dawn comic book. So Roger, Ashley, and Kris, welcome to The Geek in Review.

Ashley Dobbs 1:18
Hello, hello,

Roger Skalbeck 1:20
great to see you wonderful to join the conversation,

Marlene Gebauer 1:22
the gang, we’re going to talk about art, copyright, comic books, and AI. Are these four things that can coexist? Or is there always going to be a conflict as these pieces overlap? Roger, let’s start with you and talk about your 50 Copy run of section 101 of title 17. Comic book, talk to us about this. I mean, it is amazing. It is so beautiful. Talk to us about this, this comic book, you know why? And how did you create it?

Roger Skalbeck 1:52
That’s great, I’m really happy to join the conversation and talk about it. The origins of this go back to back in October, November, I was preparing a talk with our local CLE IP group about the kind of emerging issues of artificial intelligence and intellectual property. And so I was looking into some issues and was able to find a lot of good examples as to what was going on. And one of the stories that I discovered was the the Zarya of the Dawn example where at the time that I was researching it, everybody was excited because aI had been used to create this product, and it was registered as a copyright. Then as things go on, I find that then the story changes. And we’ll get to the evolution of what that looks like, because the author is here, which will be great to hear about. And what I read in that was a lot of it was that the Copyright Office needed better examples as to how things were created, and also what was submitted to them for that. So I was like, you know, I want to do something about this. And I thought, you know, what I want to do is actually start using some of these tools to express my own interests and kind of my own subject areas. The first idea was I was going to do a poster. So I was going to do a poster, and it was going to be the Fantastic Four factors and things like that. I thought this is gonna be great. And I started using mid journey to do all this stuff. And then I realized that that’s more of a trademark issue and that might muddy the waters as to what was going on. And I thought also fair uses that’s further in the copyright statute. Why don’t we start at the beginning and sort of the origin story. So I decided, let’s do copyright one on one, which is literally the section where all the definitions are. I teach copyright, which I’ve done since 2019. And this is a place that we look to to understand what exactly copyright is. And I’m always bringing the students back to it to say, what is the language of the statute? And how have courts done something about it. So what I did then is basically created a comic book, that’s almost every single definition of that set up in alphabetical order and put it together as a narrative and a story about the law and included the law within it.

Greg Lambert 4:09
But it’s very beautiful. Can you just describe it to a little bit for us?

Roger Skalbeck 4:14
Yeah, the and I’ll give you some copies of this too. So the best way to describe it is what I tried to do in putting this together before I’d done any image creation is I wanted something that looked like the comic books that I grew up with the Avengers, the Spider Man the kind of things that would show the heading across the top where we have the Comics Code and things like that. So I wanted something that was like in the style of of Jack Kirby and Chris Claremont and kind of the classic superheroes that I had. And it’s nice Technicolor feel to it and sort of not really low res, but sort of a simplified aesthetic. So every single image within this has a really bright vibrant color that sort of I tried to create so it jumps off the page. and, and really, really portrays the actual text of each individual definition in the statute. Well, you

Greg Lambert 5:06
did a great job.

Marlene Gebauer 5:08
How did you get it made? Like, what was the process for that?

Roger Skalbeck 5:12
Oh, great. So and this actually is listed in the back of the title page of it. So the tools that I used were, so I use mid journey for the original image generation with iteration after iteration of images. I used a lot of Photoshop and Pixlr to update it. And then what I did was, I put it together with a relatively simple program called Comic Life three, not anything, Chris has used this one as well. And it’s a, it’s a layout program a little bit like, Microsoft Publisher, kinda old school, it’s kind of simple. But what you do is you set up a layout of how many images you want on his page, then you import each of them and annotate them. And then in terms of printing, there’s some nice online print shops that do comic printing, that’s a nice full bleed approach to things and you set up the number of pages and the style of paper you want and all those kinds of things.

Greg Lambert 6:07
Very cool. I have to check that out.

Marlene Gebauer 6:09
Yeah. Does that mean that that that looks professional? I mean, yeah, I was just blown away when when I saw it in the mail.

Roger Skalbeck 6:16
Yeah. And initially, it was gonna be an eight and a half by 11 layout, which was the easiest thing to do. I printed a version, I showed it to someone. And the response was, that’s not a comic. Yeah, comic, they’re very specific size. And there’s evolution of sizes over times as to the small medium, and the US and things like that. So I wanted something that really kind of fit the mode and was of the time of the things that I was trying to emulate.

Greg Lambert 6:40
Well, you did, again, did a great job on it. I have my copy prominently displayed here in my office. So Ashley, you run the IP and transactional law clinic there at Richmond Law School. Talk to us a little bit about what kind of work you’re having the law students do. They’re at, at the clinic. And as I was prepping for this, and Roger and I were shooting some emails back and forth. I, I understand Roger may be a client, as well as the clinic.

Ashley Dobbs 7:08
Yeah, that’s right. Yeah, no, thanks for having me. I’m happy to talk about the work we’re doing. And I should preface this that Roger, who is a client of the clinic has waived his attorney client privilege, right, in order for me to talk about his project with you. So so the IP and transactional Law Clinic is an opportunity for law students, mostly third year law students, people in their last year of law school, some but some, in their second year, to work directly with clients on those kinds of matters under the supervision of a licensed attorney me and to start to bridge from the doctrinal the book learning they’ve gotten in law school to actually applying it to clients. But without being in the crucible of a law firm right off the bat. We try to work with people who can’t otherwise afford legal services. And we look for interesting projects and all kinds of realms, but we tend to work with startups, entrepreneurs, makers, creators of all kinds who are coming to us wanting to protect various kinds of intellectual property, mostly copyright and trademark, sometimes we will do have delved a little bit into trade secret. We don’t do patents, since it’s for variety of reasons. But we might let a client know and refer them to some local practitioners that do patent work. But we also end up doing a lot of the transactions related to that. So somebody will come to us and say that they want to register a trademark. And then we ask a few questions. And the next thing you know, we are forming an entity providing the guy the organizational documents and the bylaws or the operating statement, reviewing the contract with whoever helps them create their logo or perhaps creating a contract with whoever created their logo to assign the rights taking a look at perhaps if copyright protection might also be warranted, as it is in many cases, as well as the trademark work and then various other transactional matters that entrepreneurs might meet.

Greg Lambert 9:05
The big takeaway I got from that is Roger can afford an attorney.

Ashley Dobbs 9:13
I will sometimes make exceptions for people who can’t afford an attorney if they bring me a really cool project.

Greg Lambert 9:20
Well, Roger, thanks for for waiving your attorney client privilege on on this, I’m happy to do it. So Ashley, I also understand you do some other interesting work. Roger mentioned something about there’s an improv group to something that was can you tell us a little bit about that? I don’t know if they’ve waived their right yet. But

Ashley Dobbs 9:43
oh, no, but if you’re ever in the DC area in January, you should check out Elvis his birthday Fight Club. Oh, wow. They also have shows in Maryland. It’s usually on the anniversary of the The man’s birthday. It was delayed this year due to COVID. So they and it was on hiatus during COVID. So 2023 It actually I’m sorry, 2022 it actually took place in August. So they do a hilarious, hilarious burlesque show that has modeled on fight club emceed by someone who is not Elvis and he is possibly an impersonator, but might be impersonator of his ghost. But the toilet Oh, no toilet is actually figure. Yeah.

Greg Lambert 10:36
That is funny yet sad. I’m surprised th at you talk about it. Because?

Ashley Dobbs 10:45
Well, because you know, that’s my club. Right.

Greg Lambert 10:47
Right. And the second rule a year.

Ashley Dobbs 10:50
And the second reason The second rule is don’t talk about fight club. Yeah. So we try not to talk about this too much. Because Fight Club Fight Club is taking full advantage of their rights to parody. Public figures and creative works under the fair use exception to copyright.

Cool, very cool.

Marlene Gebauer 11:09
All right. So we’re going to scale it back from Elvis, and we’re going to talk about copyright again. So Ashley, the US Copyright Office last week, came out with a clarification on how it’s going to handle issues of AI generated artwork, and what is and what is not avert. So can you walk us through what the current guidelines say? Right. So

Ashley Dobbs 11:33
so the current guidelines, but by the current guidelines, I’m assuming you mean, what they said before I issued this. Yes. Okay. So there’s no mention of AI artwork, what right now the current guidelines of how you register. I mean, there is a mention of a bit not as explicit as in this guidance. So currently, when you register a work, you there’s the things that you authored, and there’s the things that you are claiming, and sometimes what you’re claiming includes stuff, elements that are not copyright a bool or are not claimable. So for example, if you, if you create an illustrated map, United States, and you start out with a map that was issued by the US geo geographic services, you know, that USGS, and that’s in the public domain, you have to identify that you’re not claiming authorship in that underlying public domain work, you’re not claiming it at all, not just claiming authorship. And you have to say include that in your description of excluded materials. And which then begs the question of what is included, and so you’re supposed to provide a short statement of what is included out. So examples of things that aren’t claimable, or the example I just gave you a public domain materials, but also to previously published material. So let’s say I’m doing a derivative work, I’m doing a revision of a book that I wrote before, I have to identify that this current claim does not include the previously published work or the previously registered material, but only the new stuff, the new elements that I’m including. And typically, the explanation is a very short description, excluded materials previously registered text, included materials, text, eight, so it’s pretty straightforward like that. And if you go to the Compendium on the copyright office’s website, it gives examples of how to do this and gives the short explanations. But it also tells you if there is work included, that’s uncopyrighted bull. So for example, short term short phrases, facts, figures, you do not have to explicitly exclude that because it’s a given. So if in your mind, what you’re putting into the copyright application is wholly yours, and nothing is excluded, because it either falls into uncopyrighted elements, or because you are claiming it. There, there was never a place to write a dissertation or a lengthy explanation of what you’re including and why you’re claiming it. And why until they asked you about it. And they’re not going to assume anything. It’s not as if you don’t exclude something, or if you don’t claim something, they don’t make any assumptions about what you meant or what you intended.

Marlene Gebauer 14:23
Alright, so and then with the new clarifications, have they struck a balance? Or are we still unclear about what they’ll do when, you know, Roger goes to file copyright protection?

Ashley Dobbs 14:36
Well, like, you know, we’re still taking a look at what they issued and sort of analyzing it. At first glance, I feel like what we included for Rogers claim is within the spirit of what they’re suggesting, in that we disclosed that AI was used as a tool and we identify the elements of human authorship. They have some space cific requests were like don’t name the tool that you use just say that you used an AI tool we did. But I don’t know if that’s material. And if they’re going to make us revise it as a result of that. The other thing is, is it doesn’t really help where you maybe have a borderline question of whether something could be claimed or should be claimed as human authorship. And, for example, there are circumstances where the question hinges on how much human interaction there was with the AI tool. And I don’t feel like that that’s adequately addressed in these descriptions, it doesn’t help with gray areas. And the box on the online form is pretty as well as the paper form is pretty darn small. And there’s no way that I can find it provide an attachment, except if you go into the comments and like, see how many characters that allow you to put in there or I guess, say, you know, contact us for more information, or what have you. But we’ve that the key tenants, I think of what’s in the new guidance are, identify that you’ve used AI, identify the the human parts that you the human generated parts that you weren’t claiming, and identify the parts that you’re excluding. And I think we’ve complied with that. But I still think it’s possible we

Greg Lambert 16:15
can you do what I did in law school and just shrink or enlarge the font to fill the space that you need? They allow you to do that

Ashley Dobbs 16:24
with Matt, when I wish that we could, because that is no, it’s just one of this online, their online form is not user friendly. Yeah,

Greg Lambert 16:35
there seems to be agreeing with you. I was just gonna

Roger Skalbeck 16:39
say who had two things. One, it’s very user unfriendly. And what we did when we were trying to put this together is actually be descriptive about what went into the application. I my publication date, I think came out the day that they issued a revised guidance on Kris’ registration. And then a little bit later, they issued this new stuff in the Federal Register. So I didn’t have the benefit of it. But I was guessing where things were going. And I kept having to revise it, because it was like, this number of characters didn’t fit. So it was kind of it. You know, in the moment editing process, there is some degree of optimism that the copyright offices started a modernization process. And they’ve had think, typically twice a year webinars where they show off kind of where they’re headed. They have a number of people and a variety of interests represented there. It looks really cool. This is kind of your typical government project, where the end result I think, actually will be very admirable. The process by which we get there is probably about 24 months before there’s a new system. If I understood their explanation correctly,

Greg Lambert 17:50
Well good. It’s about time they joined the 21st century, right?

Roger Skalbeck 17:55
Yeah, we’ll get there.

Greg Lambert 17:56
Well, Kris, I know we’ve kept you kind of hidden in the background while we’re talking. But we are really excited to have you on because of your experience with your book, Zarya of the dawn, and it’s been through, I would say, like this pinball machine with the US Copyright Office where it was initially granted protection, and now they’ve reversed themselves. And so I mean, it’s this back and forth. But let’s start off at the beginning. And have you talk about Zarya of the dawn and the way that you created the story, you laid it out the artwork, and you know, what was the creative process that you use to create this?

Kris Kashtanova 18:39
So I wrote a story before AI became such a popular thing. And actor, there was no way for me to generate those images with AI because they were no tools that were available for public. At the time. I think there was a waiting list to get open AI DALL-E, and I was on that waiting list for many months. And I only was able to use it when they went public and you could actually get it for mining. So I actually when I was writing a story, I wasn’t aware of generative AI tools for images and that’s that’s going to become such a big deal. It was 2021 very deep into the pandemic. I live in New York City. I live in Manhattan. So everything here was very, like, streets were empty in 2020. And in 2021, things kind of started getting a bit better. But still, there was a lot of cases in the city and I felt quite isolated. They didn’t See my family for a very long time, and I lost my grandma who raised me. So I lost my best friend, and then another best friend to counselors, it was a time putting Eve. And it was a time of feeling isolation. And I also was unemployed for quite a long time by then, because I am a professional photographer by trade. And I used to photograph sport and fitness and yoga. And my job required me to travel, or my clients to travel to New York, or to at least being able to meet people in person, which wasn’t possible anymore. So unfortunately, all of my clients even though they tried very hard, they couldn’t really have any photography taken. So I was unemployed. And it was really hard. I was living on savings. My friends were helping me out immensely during this time. So I was in a very dark place. And then imagine this August 23 22. And I see this images climbing on my Facebook feed of people who weren’t particularly popular on social media, they didn’t have a lot of following. Because I remember that to get the dialer, accounting had to be like, either famous or an influencer or an AI researcher. But those were people I just knew, my friends who weren’t famous in any way, and I see this incredible images. And some of them were writing, this is mid journal. And my first reaction was like, Oh, wow, how did they get an invite? So I asked one of my great friends, how she got an invite, and said, Oh, anyone can join. It’s not very user friendly, but you just go in the discord. And create yourself an account. And this is what you do. And it’s just sent me the website link. And that completely took my pain away. Because I was struggling. And also during that time, I had dental surgery quite quite like half of my face was swollen. And unfortunately, dentists didn’t give me any fun medication. So they just said, then you can take it, a lot of it and I was in so much pain and distress. I just wanted something to take my mind off of my losses, my surgical pain. And then I discovered me journey. For those three days, I generated pretty much everything I could think of. I didn’t, I wasn’t making a story, I was just generating everything on my mind. And I live in New York, I really like Central Park and T’s and nature. So I generated every corner of Central Park. Then I generated an organization, I volunteer with backpacks for the street. But the images were so ugly that I was so ashamed to show them to my friends who were volunteering with me, because it wasn’t very good at generating humans in such a great detail or especially few few months and inaction, right? But this was my three days and I exhausted by imagination for all the images and I thought, okay, now I generated everything. What what’s next. And that’s when I remembered the story about a year. And I thought, oh, that will be actually great because I tried to illustrate that I tried to use I take self portraits, and I was also learning 3d To slide blender and cinema 4d. So in 2021, I tried to illustrate my story with 3d and self portraits. So I was just combining 3d worlds I was making in blender and then taking photos of myself and putting myself in and I tried to like turn around so you can’t really see face. It’s just like my character who was always available to pose for me because it was me. And I’ve had about 200 images because I did it every single day. But they weren’t climate vocal correlative because they were in close ups. And they will know like a vast scene and a tiny, tiny human in the middle. So it’s very hard to make a comic book out of such similar images. And then I was writing my story along with every image. So I already had the whole story, and it was laid out. And the idea was that every single day, I would create a different world. And all of those worlds were connected to my emotions. So they were like, the worlds themselves were the way I felt on that particular day. And the world itself would help me to resolve that feeling. I take a lot of psychology courses, it’s my biggest spat with psychologists, I was like, I’m going to share everything I learned so far, that helps me a psychology deuced through this narrative, and it was like a long, long, long, long story, your long novel. And then when y happened, I was like, Okay, I’ll try to make one first chapter. Because the story is so long, so I had to kind of take just a tiny bit of it. Yeah. And that’s how Zarya of the Dawn happened. And this is how I started generating those images. Every single day, I would write a one page, and I would share on social media. And I noticed how AI community was really supportive. And just the reaction was so positive, because I shared that and AI community groups. My love, people love it so much. And then I thought, well, that it will be great. My first thought was, well, maybe if I make the entire karma book, I can print it and offer it to people donation based and raise some money for the charity. I am volunteering, right, because it’s very small. And they always need my net. And I did.

Greg Lambert 27:23
It’s really interesting, because I think this is a story that you hear a lot, especially with the artist is that, you know, a lot of the greatest creations come from some, unfortunately, amount of pain. And that’s the way that they’re able to focus. It’s why artists put out such great work, even even when, you know things are terrible with them. Well, let me ask you one thing about the online community, because I’ve got a few artists in my family. And when I talk about AI generated art, sometimes I get a negative reaction. Did you did you have people that were were coming from it from a different angle that was seeing it as a threat rather than another form of expression?

Kris Kashtanova 28:08
Oh, yeah, it’s actually a very current topic.

Marlene Gebauer 28:12
I’ve seen it in the news quite a bit.

Kris Kashtanova 28:15
Yesterday, on Twitter, I was compared to Hitler. And took ribride ai images. And it’s quite painful, because actually, my great grandfather was at concentration camp. Oh, my God, he did survive and had a very long and very good lives, she ran away, but my good digital year, such thing and also from people who don’t know me, and seeing this is still going on, because when I started in August, I was kind of a little bit shelter because I had my Facebook groups of AI enthusiasts and it was not that big. And people were just being excited. When they started sharing on my main page on in Facebook, like a personal page. All of my friends photography is kind of divided into two groups. First one, really wanted to learn it. And second one was completely against it. It almost felt there was no middle ground. And you know, I come from software engineering background. I worked as a software engineer for 10 years. So it wasn’t just my artistic side that was excited. It was also a computer scientist. And because I have master’s in computer science, I worked as a programmer and you know, when I was studying, it was at 2000 to 2007 I did my masters. I saw that AI we did have a subject That was called AI, it was the most boring subject that

Marlene Gebauer 30:05
was happening.

Kris Kashtanova 30:09
The person so uninteresting to me. And I remember, I always wanted to get get good grades. So basically, pretty much I know onto it, passed an exam. And I only remember the pic and had artificial intelligence and writing on that. And it was no pictures, just text. And then it was it was large, it was huge. And now when I think about it, they think, Oh, if I could have a time machine, I would read it, I would read every single word as I am interested where things were back then. But right now, it’s so exciting and so amazing. But I think computer scientist in me was rejoicing. I was like, Oh, wow, this is happening. And because I also have an understanding of algorithms and how it works and how the program, how it operates. It blew my mind on so many levels. And to be honest, I am quite sad and that it keeps happening the harassment and bullying and some people say, Oh, you can call harassment or bullying, just if people have different opinion. But that’s not

Greg Lambert 31:32
true. That’s happening. That’s not

Marlene Gebauer 31:34
That’s not what was happening there. I mean, that wasn’t a Hitler it. Yeah,

Kris Kashtanova 31:38
right. Some death threats. And not just to me, AI comic book writers, but a lot of death threats. So I think it kind of crossed the line at this point. And I feel like as different tools come out, this is what happens, things were pretty. Okay, it was not perfect. When it was just me journey. As soon as Lenon said he needs to do self portraits. That was the first time the level of harassment was so high. I honestly didn’t know what to do. i It was around before Christmas. And it lasted from like few weeks before and then all the way into Christmas. And I was thinking, well, this is a very uncomfortable situation, because everyone knows my home address because of the copyright office who revealed it to the entire planet. And I am being harassed and so in such a way so I really follow the beat strange about it. But you know, there is a positive side about it. Because one day, I get this beautiful package in my mail. And I open it and I couldn’t believe it. Reuters sent me a car cart by mcpaws about a like Edgar Allan Poe. And she also wrote me about Edgar Allan Poe because I didn’t know that he lived in Bronx and I also didn’t know he was being corrupts when he was alive. So I was like wow you know, it’s it’s every every model has two sides. So yes, it was harassment but also, so much kindness so much support so many new friends and connections. Also some people who were quite extreme about AI, now writing me and say, Well, don’t tell anyone I use AI but can you tell me how to

Greg Lambert 33:51
do all the cool stuff you’re doing without being referred to? Is some person

Marlene Gebauer 33:57
they see now you’re the expert people are going to use so you know, speaking of that, I know that that sorry of the Dawn was you know, we talked about use the you were using mid journey to generate that artwork. But now your newer creation rose and Nygma is using stable diffusion and a different AI generation tool with a combination of your old pencil drawings. So how is this style different from what you did on sire of the dawn?

Kris Kashtanova 34:27
So well, I am an artist. I make a lot of handmade stuff. I make clay mansions. One of my biggest passion and dreams is to make a sculpture out of garbage materials like recycle materials, and take that creation to Burning Man. That deal Sam and my apartment.

Marlene Gebauer 34:54
That’s cool. I just I just read a fiction book by ruse as Becky that dealt with art Made of garbage. So that’s just the thing. I need to read that and I’ll send you, I’ll send you, I’ll send it to you how many

Kris Kashtanova 35:06
things you can make. I saw, I saw this incredible artist in Portugal he makes out of like metal, or cars and stuff like this bind is a bit more just it just like everything was thrown away. I had one sculpture made of boxes, like milk boxes, and like stuff like that. And I made my mind out. There was like, be construction, like little houses. And then it had lots of light in it. And, and then I made the dragon out of CDs. I got some old CDs from eBay. And then I cut them out, like I made triangles out of them, like 1000s. I did have time during the pandemic, as you can see. Right. So I’ve got spent weeks catching those CDs. And then I stick to them. I made a sculpture out of Bobby Amercia. And they made a great nice dragon out of this eerie descent like material. So this was something I do. I am always dreamt to be good at drawing since I was born. And I was born in North Caucasus and USSR. And my family, my family was so poor, but they did try to give me an education. And there was an art school in our village. And one day I went there with a friend in secret I just sneaked in. I was like, eight years old, and I just sneaked in. And my friend told me all they want notice extra child. And rock by school, it had like easels and paints and watercolor and everything. And I it was the happiest day of my life. And then when I got back home, my grandma was completely mad. We call us. She lost me for the entire day I was away the entire days. She didn’t know where I was, because there were no cell phones. We didn’t know how that’s gonna happen. And she was freaking out. And she was so angry. And she told me where I went. And I said, Well, I went to art school with my friend or learner. And I want to go to this art school. I want to draw that solo one. And my grandma, she was a very kind woman, even though she did get really mad at me, but she was very unkind. And she said, Yeah, let’s try to make it work nice. And unfortunately, she couldn’t, we didn’t have mine. So I couldn’t go and the school asked us to pay for the entire year. And it was already like, and the semester so my grandma was like, we don’t have money like that, oh, we could have paid for maybe one once the one that is left, but not the entire year. But because of a lot of people did that school just to get like a certificate or some kind of thing that they could start with later. And it was like, eight years long school, it was like a proper art school. It just didn’t work. And I remember I was so sad. But also I knew that she will be sad. If I show that I’m so sad. I was like, Okay, I’ll just focus on earning money. Growing up getting a job earning enough money and going back to art school. But the problem was that my entire life I spent doing math and physics, and I didn’t have much time to practice. So my drawing skills stayed the same as it was when I was eight years old. And for many years that I did try to take classes, I did try to get good at it. But you know, when you do something commercially like photography or something, you just don’t have time to develop another skill to take this time out of work. So it always stayed on the back of my mind. Oh, and I was good. When I was eight I was really good. But now I’m not I couldn’t be in commercial or not.

Ashley Dobbs 39:27
Yeah. Well, last summer I taught a class on the art of Burning Man and intellectual property. And I have been part of teams that build art at Burning Man and installations at Burning Man. So if you want to talk about that offline, let me know. Yeah.

Kris Kashtanova 39:48
So to continue this, I only want to mention that even though I still haven’t developed my drawing skills because I didn’t have Time, I do still have ideas of how I want those things to look in my head. So I know how I want to eat to be drawn. And I can draw in some way. But it’s never the vision I have in my head, the idea of having my head is very hard to make. With me drawing, I can make it through photography, I got pretty good at Photoshop, and taking photographs and making like collages out of it, to show it in such way. But to draw in such way, unfortunately, it probably I probably could, if I did it full time for maybe five years. But I’m 37 years old. And I have other things that responsibilities. So I can just drop everything and spend this time we’re learning how to draw. So for me, I was looking for tools, how to do that. And you know, first, when when they just grow during the AI community, the journey I tried, there was a way to draw, put the drawing into the journey and see what happens. But the results were very good back in August, it was nothing like my drawing, it would just do something or that it would almost use it like in order to take my drawing, take one color out of it and somehow arrange the image wasn’t controlled in any way, it didn’t look like my drawing. So there was no point of me trying to draw something. But few weeks ago, stable diffusion, which it’s possible to use completely free on your own computer, released this control net, which is a way to direct the modal. And you can do a lot with it. For example, you can, you can generate something like a person, and you can make a doll, that kind of a puppet shows that bows. And it will be exactly like the policy made, you can put an image like a sketch. And then you can generate something that resembles your sketch and you can direct it if you can say what exactly you want with your sketch as a reference.

Roger Skalbeck 42:29
It’s interesting here to hear you describe the process and the advances in the Tools sort of adapting more to your both your intentions and the direction you’re going. And most wonder if there’s an analogy to music, which is if you use something like auto tune, or reverb or you know, production, you know, facilities, you say, Well, I can’t sing. But I have lots of techniques and lots of technology that makes me a better singer. And if I were to apply for copyright for the sound recording, I wouldn’t have to disclaim and explain all of the audio, you know, amplification and sort of tuning techniques that I have, it seems odd that like there’s this approach to saying, Well, you need to do all of these things in AI, and especially mid journey three and four and a little bit into five, well, you don’t have a lot of control over it. So we’re not going to really give you much there. But it’s nice to see stable diffusions models, and evolving into a way where you both have more control. And also your vision can be something that’s not accidental, but intentional.

Greg Lambert 43:37
And I think one of the things that you’re seeing and and I’m kind of realizing this firsthand, is that being able to use the AI tools, at least as they are, you know, in March and April of 2023, that the skill set to be a great writer, or a skill set to be a great artist don’t necessarily run parallel lines to what it means to use these tools to create great writing or to create great drawings or images. It’s a different part of the brain, I think, is probably still on that, that right side, artistic side of the brain. But it’s different. It’s a lot more about being able to describe what it is that you want to do rather than just do what you want to do, which a lot of times I think with, with artists, you know, they’re trained in order to kind of take a vision in their head and then take that and make the make the art out of it. And not necessarily, I don’t think they could necessarily before they start, describe exactly what it is that they’re going to output. It’s kind of a process in to bring it back to the law. The biggest obstacle that I’m seeing for or lawyers and AI tools to advance their ability to practice to streamline some of the things they do is they kind of they’ve, you know, they’ve been this is a different way, they had to retrain their brain to think differently about how the tools approach. And to me, it was very frustrating. I even had a tweet out this week that I was a little frustrated, because the lack of creativity and imagination that I’m seeing is really hampering where I think people could benefit from it. But you know, I might, I might just have to tone that down a little bit on my side, because it it is a retraining. And if and if that’s not a part of the brain that you’re used to relying on to do your job, it’s it makes it really hard to kind of figure out how can I leverage this great tool to make what I do even better,

Ashley Dobbs 45:58
as you were saying, and talking about and legal technology to take that risk. And Roger and I have talked about the the the in we talked about with our law school, too. And we’ve added courses about thinking about design and systems thinking, because you’re thinking about the process,

Kris Kashtanova 46:10
I’ll say, we’ll even though I already know those guidelines, say a bit vague. But to be honest, this is the first time in history such a revolution is happening. And I think because the two sides evolving so fast, like, literally every minute we are talking, I’m pretty sure I will check my use after the stock, and they will be new tools and new algorithms and new stuff. So I feel the best we can do right now is to educate the cooperator first because the job isn’t to learn every AI technology, then they probably don’t have an AI consultant. I mean, I would love to be an AI consultant for them if they need to. But I feel like they already have added other tasks. And if we can lay out those tools for them in a way like the lawyer would explain it because I wanted to do it by myself, because I can explain very well how I do but I do it as a computer scientist, what I realized is that what seems very obvious to me, isn’t obvious to maybe lawyers. So that’s why I requested I wrote on LinkedIn. And I said, I’m looking for a pro bono lawyer to help me write this cover letter. And I was really lucky because not only I got the pro bono lawyer, the lawyers who were working with them, they were interested in tech, and I spent hours explaining technology to showing all the new things I’m using in my process. And they spoke the same language, because it’s one of those things that I don’t know, it will keep evolving. And artists also need to have a way to understand what is copyrightable and what is not. And right now the guidelines is just so vague. I personally feel that those images I created in zero, Dawn, even though I did not draw them. It wasn’t like a slot machine. It was the prompting the process of me those prompts were very long because it was old version. So there was creativity in me making those prompts and directing it. Also, it again, it wasn’t just random output. I was directing it. I didn’t get like cats and dogs and clouds and food. Now I got exactly what I wanted and how I wanted that through modification of those prompts. But what I understood is just difficult for them to grasp it. So I want to show all the possible technology right now. And maybe one day, we will have images just text to images, copyrightable in some way.

Roger Skalbeck 49:16
What’s interesting hearing you describe the process, Kris, there’s a class action lawsuit that’s been filed in I believe the Northern District of California. There’s three artists representing it. They’re suing stabled diffusion majority. And I think DeviantArt. And one thing that’s in kind of the explanatory part of the complaint is they’re they’re trying to describe this as if the process of using AI tools was like querying a database. They said, You’re just you’re searching for things and they’re there and you’re getting output. But it’s I think that that really misses the mark in terms of both what was happening even for the versions of those tools at the time, but also, just generally what’s happening bidding. And also the the other piece of it on the stable diffusion model is that the the large corpus of information that is in a lot of these tools is, you know, multiple, multiple terabytes. And it’s not that you’re querying to get a copy of something like you are from Google image is, you’re creating something that’s an idea of an apple or an idea of a tree, and how it’s shaping that in terms of kind of morphing things together. Yeah,

Kris Kashtanova 50:27
yeah, it didn’t exist before. The other way for you to create exactly the same moment, as I did is the process I described in my copyright letter for Rose Enigma, because I revealed to you my seed. And that’s a very big number. And I revealed every single setting I use. And the thing is, it’s like, it’s it’s quite similar if I was making a photograph, and I would reveal to you the location them all the time of the day, lighting, and all the settings I used in my camera, you could potentially create the photograph just like mine. But normally, lab photographers don’t do it. So it’s not like I’m planning to review seeds and everything for everyone. But I just wanted to show that if you put all those parameters, the same prompt that I created, in my mind, you will keep getting exactly the same image, you can also recreate that with your stable diffusion. So it’s not random. It’s not like it came from some something out there. And there are there is limited amount of limited images you can create, it’s just the number is so large, because you can change a brown just one letter and a different seat, I was do something else. But it is a limited number. It’s not indefinite if you kind of think of math and stuff like this, because the prompt also has a limited number of tokens, you can use it just so big that probably I don’t know, it’s not teaching.

Roger Skalbeck 52:11
And one thing that’s interesting in some of the guidance that seems to come out, or one way to interpret is that there’s a little bit of a requirement of show your work, like you know, tell us what you use tell us the tools and things which I really appreciate the approach that the attorneys that Morrison Forrester took to VT be very deliberate about that. But you know, we learned from, there’s this case that went to the Supreme Court involving the white pages, so telephone directories. And the basic premise to that is, copyright does not come from human labor, it comes from creativity. So it’s not that you put more work into it. So you therefore get a copyright, or you put less work into it and you don’t. So there’s this what’s called the sweat of the brow doctrine. And I think there’s, it’s an interesting tension between show more of your work. And well, the more work you do doesn’t necessarily mean you get better protection,

Greg Lambert 53:04
and Ashley and Roger, and, Kris, as well, going forward, you know, how should people who are creating, doing this and are looking for copyright protection? Because I imagine, you know, Roger, you mentioned auto tune, I imagine there was probably some talk at the time that things that are auto tuned. Maybe are not artists are not at a level. Or you can you can look at to Photoshop, I know when was an Image Comics are one of the comics that started using actual computer programs to draw their comics and color. Their comics was that somehow less artistic than, you know, the pen, the pencil and pen and ink. So I imagine we go through the stages as the technology advances. And there’s a change in the way that things are produced. And again, like like you said, Roger, it’s not necessarily about the amount of human labor but rather the creative process that’s involved. That is what is the reason that IP law is built into the Constitution as we want to have this incentive for people to create? How do you see this evolving? You know, as we go forward?

Roger Skalbeck 54:32
Do you want to go first? Do you want me to?

Ashley Dobbs 54:35
Well, I mean, I was I’m still thinking, but one of the things I would say is like this goes back to that, and you will remember the name of the case, Roger, because I still can’t remember the name of the case about the camera. And whether photographs were copyrightable.

Roger Skalbeck 54:49
Right? This is bro Giles. So this is the this is the question of whether or not somebody should be granted protection for a photograph of Oscar Wilde because the claim aim at the time was that the constitutional origins are that it protects writings. And a photograph is not a writing. At least that was the argument. And thankfully before they actually put it in Section 101, of title 17 of the US Code, they said no, actually, this is a writing, okay, we can get there, when it’s not even that big of a mental leap to go there. Interesting.

Marlene Gebauer 55:24
Roger, you’re going to launch a course in 2023, called comics in the law, where you’ll work with your students to actually make comic books using AI, Creative Commons resources and their real creativity to create access to justice thing, comments, that sounds like something that I truly I would want to audit. So yeah, what’s, what’s the goal of this course? And what value do you think students and others might get out of enrolling in the course?

Roger Skalbeck 55:52
Fantastic, I love I’d love talking about this. And I’m so excited about well, this is cool.

Marlene Gebauer 55:58
I thought I thought about this, I thought about this in terms of the comic that you sent in terms of different ways of, of learning. And I mean, I thought about like, I’m going to show this to my kids and see if like they can ingest that more quickly. So the fact that it’s a graphic novel?

Greg Lambert 56:15
Well, since you mentioned that, Marlene, I want, I want to just point out that when I was a kid, the thing that got me most interested in reading was comic books. You know, I was not a great reader. But it was something that I picked up, it was something that I had an English teacher who let me run with that. And so I mean, it’s there’s great value in having different styles of teaching, a topic that it can based on the students that you have. So sorry, Roger, I didn’t mean to answer that.

Roger Skalbeck 56:51
No, that’s, that’s great. No. So I’ll, I’ll give a couple of examples of the kinds of things that I think are both inspiring to me, and then talk a little bit about where I think I think students should go, and we should be thinking about this. If you look at something like law tunes, there’s law tunes and law tunes to it’s from somebody who lives in the Netherlands, at The Hague. And this was developed as a way to create a comic book to help kids understand the law better. And in they then I talked at some length with the person who put this together, it’s a fascinating project to look at. And she was very specific that they wanted to think, how do we make it so that was more understandable, and accepted in terms of like a, a foundational element of society. So it’s a really interesting thing there. And there’s both a really cool comic out of it, and a fair amount of academic research that’s gone into it. And another example of things that I’ve seen that I’ve been really inspired by are materials that come out of the Harvard access to justice lab or a to J lab, what they’re doing is they’re using things such as one in two page comics that explain to people how to respond to a summons. And instead of putting it in dense language of long narrative paragraphs, it’s a four to six panel comic that shows somebody getting a letter, opening it a specific number of days that you have to do something with it and putting it in the envelope. And sitting behind all that they have empirical studies that really measures all of the all of the outputs and all of the achievements that they have. I want to take our students and jump to the end of it, which is tell the story about access to justice, tell a story about a legal process or something important to society, and develop the narrative storytelling features of it so you can get there. My counterpart is a big fan of there’s a couple of books by Linda berry that are about making comics. And Scott McCloud does some as well, that are the narrative style of pen to paper. So he’s anybody can be an artist will do that. I’ve been inspired by the AI tools. It says anybody can have an idea. And there’s a various number of tools that you can use to get there. So it’s going to be a combination of all those and there’s two requirements come with an open mind and create one comic as an output that connects to justice.

Greg Lambert 59:19
Yeah, definitely want to audit this class now.

Ashley Dobbs 59:22
Yeah. Me, Too!

Greg Lambert 59:25
Kris, we ask all of our guests will we call our crystal ball question. And Roger, Roger, feel free to chime in here as well, if you want. What do you think I know, we’re just at the beginning stages of this combination of humanity and AI when it comes to creativity. I mean, just in Kris’ story between August of last year and now there’s been this just enormous amount of change and like they said, we go check our our email, there’s going to be something new and They’re this afternoon that wasn’t there this morning. So where do you think, you know, when it comes to the issue of what will deserve copyright protection? Where do you think we’re headed in the next, you know, two to five years? nationally? I’ll let you start off if you don’t mind. Oh, sure.

Ashley Dobbs 1:00:19
I’m having to start. You know, I think ultimately, that the Copyright Office is going to have to recognize that these are tools that people use for creative expression, just like 100 years ago, the camera was another tool for creative expression. Just like auto tune, let Roger mentioned, you know, there are so many examples of a director putting together a movie telling and telling an actor how to deliver the line, there are just so many ways where we recognize creativity that uses tools that is not simply pen on paper, that I think that it’s inevitable that they’re going to have to recognize this.

Greg Lambert 1:00:56
And Kris, what do you where do you think we’re going in the next two to five years?

Kris Kashtanova 1:01:00
Is it two to five years?

Greg Lambert 1:01:02
It may be it may be too far,

Kris Kashtanova 1:01:04
you know, to five months, I’m thinking about it every day. And you know what? I’m known for first being too optimistic. In general, that comes from my grandma, who went through a war famine, and she was still so positive. So and she raised me. So it’s like, we lived in a conflict zone. And I still was like, Yeah, this is a bombshell. But we will go and have a great fun and like a shelter. So you know, it’s like, it’s just my attitude, even though it is a very, like difficult time for AI creators right now. But you know, what they think? I want to believe that, first of all, this technology has a great potential to cure disease. One of my biggest focuses on AI now is how can we use AI to fix climate change? Also, accessibility, the fact that they created this comic book was because my younger brother has special needs. He loves comic books, but it’s very hard for him to read the he has very skinny ADHD. And I thought, I will create a comic book that he can actually read, because he always told me you read books, you shall write a book. And once I asked him, I said, but you will you read it? And he said, No, but I want you to be happy and to make it. He’s 16 years old now. And I thought, well, he can read the comment book. So I made this Corman book, it was translated on over 30 languages, including Greek, which is the language my brother speaks, he’s my half brother. And I dedicated it to him. So for for people like him, it’s accessibility. And also I creating images, not a lot of people can draw just due to the physical limitations. That’s why Rosa Nigma isn’t going to be my last. So for me, it’s also accessibility. So I feel like in maybe five years, this thing will become a norm. And hopefully, the harassment wave also subside, because I don’t know if you’ve heard but just few days ago, Adobe launched Firefly, which is the texture, texture, image generation, and also font generation. And they train the model. And the model is trained on royalty free content they have license for. So it’s like the most ethical model that exists right now. And Adobe is very focused on creators, so they definitely will compensate every creator involved in this. It’s one of those companies who really cares. So I feel it’s just the start. Yeah, they will be more companies who do it in the same way. So the, the ethical part will be resolved. And yeah, it’s, I feel like it’s, it’s a beautiful thing. They definitely will be like growing pains and stuff like this, but hopefully we’ll overcome it. We are creatives, we are adaptable, and in five years, it will be just just a thing that we use to make our life easier.

Greg Lambert 1:04:34
All right, Roger,

Roger Skalbeck 1:04:36
I was gonna say I’m not gonna go into detail about predictions. But one of the last things that Kris mentioned, I think is something we’re going to see play out, which is companies that take an approach to incorporating tools where there’s a greater amount of clarity around what is allowed to be used. I love this idea of Adobe’s Firefly where they’re training it on the Things that have a very specific risk profile to them, which is nearly zero or things they’ve paid for. I’ve heard from somebody who works in the music industry where they think that there may be a direction that happens with that as well, where somebody owns the music rights for, you know, a corpus of materials, you could kind of play in the space. And then what you’re doing is you’re interacting with things they have license. It again, is a question of what’s the ownership piece of that is a contractual thing that runs to them as a subscriber or not. But I do think that there’s going to be a movement in the direction of maybe a little bit more clear, and sort of rights to the particular elements that are underlying these creative tools.

Greg Lambert 1:05:44
Very cool. Well, I’m excited to see what see where we’re headed on this.

Marlene Gebauer 1:05:49
So thank thank you all for coming on and talking to us. Kris, how can people find you and your work online,

Kris Kashtanova 1:05:57
I publish a lot of free educational material, I make eye tutorials. And they make explainers and all kinds of things. And it’s on my website, which is very easy to remember, it’s Kris.art. Also, I’m quite active on Twitter, where I feel like my community who uses AI is growing and thriving.

Marlene Gebauer 1:06:24
Roger and Ashley, other than your professor profiles at the University of Richmond Law School, where should our listeners go to find you online and learn more about your efforts,

Roger Skalbeck 1:06:33
I continue to be on Twitter, I mostly at this point, a reader less so of a contributor. But by all means, I find that a place that that still has opportunities for engagement. And certainly my professor profile, I’d love to connect with somebody I’d love to think about, you know, publishing this work more widely and thinking about new opportunities for developing especially educational materials and things like that.

Ashley Dobbs 1:07:00
And for me, actually, our Dobbs is my is my name on LinkedIn, or my professor profile. Those are probably the best ways to find me. The university also maintains a YouTube channel where they have professors giving short explanations of issues in the news of issues and law issues in the news, and it’s called the synopsis. So you can also find me there explaining things like why Cardi B can’t get a trademark for Oh, cool. Well, right now,

Greg Lambert 1:07:33
YouTube’s could be could be so exciting. Good. Yeah. Thank you. Thank you all for being here.

Roger Skalbeck 1:07:42
Thank you so much. Bye.

Marlene Gebauer 1:07:44
And of course, thanks to all of you for taking the time to listen to The Geek in Review podcast. If you enjoy the show, share it with a colleague. We’d love to hear from you. So reach out to us on social media. I can be found at @gebauerm on Twitter,

Greg Lambert 1:07:57
And I can be reached @glambert on Twitter. Or you can leave

Marlene Gebauer 1:08:01
us a voicemail on The Geek in Review Hotline at 713-487-7821 and as always, the music you hear is from Jerry David DeCicca

Ashley Dobbs 1:08:11
Thank you, Jerry.

Greg Lambert 1:08:12
Thanks, Jerry. All right, Marlene, I will talk to you later.

Kris Kashtanova 1:08:15
All right, bye so

Unknown Speaker 1:08:26
Hey, take five seconds to back. Devils backbone. Devils back

Legaltech Monitor is a one-stop source for legal technology news and information.
Copyright © 2024. Robert Ambrogi. All rights reserved.
logo