Engelberg Center Live!

GenAI & the Creativity Cycle: Creativity, machines, and the heritage commons — What collaboration opportunities are there?

Episode Summary

This episode is the “Creativity, machines, and heritage commons - what collaboration opportunities are there?” panel from the Generative AI & the Creativity Cycle Symposium hosted by Creative Commons at the Engelberg Center. It was recorded on September 13, 2023. This symposium is part of Creative Commons’ broader consultation with the cultural heritage, creative, and tech communities to support sharing knowledge and culture thoughtfully and in the public interest, in the age of generative AI. You can find the video recordings for all panels on Creative Commons’ YouTube channel, licensed openly via CC BY.

Episode Notes

Nitcha Tothong (eleklekha artist collective) moderating a conversation with Sasha Stiles (Artist), Max Sills (Midjourney), and Sarvesh Mahajan (Crowell & Moring)

Episode Transcription

Announcer  0:03  

Welcome to engelberg center live a collection of audio from events held by the engelberg center on innovation Law and Policy at NYU Law. This episode is the creativity machines and heritage comments. What collaboration opportunities are there panel from the generative AI and the creativity cycle symposium hosted by Creative Commons at the engelberg center. It was recorded on September 13 2023. This symposium is part of Creative Commons broader consultation with the cultural heritage creative and tech communities to support sharing knowledge and culture thoughtfully and in the public interest in the age of generative AI. You can find that video recordings for all panels on Creative Commons YouTube channel licensed openly via CC BY


Nitcha Tothong  0:53  

Hi, everyone. So this is the sixth panel and we are gonna talk about creepy machine and heritage come on the more coloration opportunity are there and my name is Nitcha Tothong. I'm also from the artist collective electric that you heard from my partner earlier in Chicago. And I'm an interdisciplinary bank bank of Bonn Berlin base and I work with technology and mostly mainly related to visual. And we have here our panel such as style, such as a poet, language artists and AI researcher whose work it mean to be human in nary post human era pioneer of algorithmic authorship and blockchain poet and author of hybrid poetry collection technology. She became the first writer of AI power and literature of major auction house and has been named one of the top artists shaping the digital art scene. Otter honor includes future art award the other woman page shot list and nomination for forward price, push card price and best of the net. Also the co founder of digital poet key point the collective and worse force. She served as poetry mentor, mentor and humanoid and Dr. B now for the since 2018. And also we have Mac sales meet journey for me journey. Mac sales is general counsel of military and operate open at a grocery store with a boutique law firm that does generals consoles, service for AI company and artists. Before that, he leaves Google Open Source LEGO Group. And lastly, we have us sandwich Mahajan from credible and morning LLP. He's a lawyer and with the law firm of Colville morning focusing on technology transaction among other things, he advise technology company on intellectual property and that our rights and responsibilities. So it also has an undergraduate degree in South Asian study and University of Pennsylvania and an interest in how South Asia cultural heritage is preserved. disseminate and transform so welcome all the panels so I would love to start with how Autolite how you interact with AI in in your creatives all like the AI from the collaboration perspective in your in your day to day life and practice. Yes.


Sasha Stiles  5:07  

Well, thank you so much for the introduction. And just first of all, thanks so much to everyone, the panels have been fascinating today, I've really learned a lot. And I'm leaving with a lot of food for thought. So thank you, Creative Commons for organizing this. So as Michelle said, thank you again, for the introduction. I'm a writer, as well as the language artists. So I am look mostly using generative text tools using large language models. And for the most part, I'm using them as a co author, I'm using them as a collaborator. And to give a specific example, I've actually written I've written lots of poems now with AI. But I've started in about 2018. Writing with GPT, two, which now feels kind of ancient, but I started using sort of an off the shelf version of GPT. Two, I quickly realized that it wasn't quite what I wanted, in terms of its output its voice. So I quickly, you know, got over my intimidation as someone who comes from more of a language, and linguistics background, and I kind of taught myself the basics of what I needed to know to start fine tuning those underlying MLMs. And basically started to take all of the poetry that I've been writing, you know, for years and years, with the intent of publishing a more conventional book, and use all that information, used all those columns, used all my associated research notes, and all the things that had been sorted, inspiring me, and that I wanted to kind of reflect in my poetry, I turned that all into a training data set, and use that to basically create a bespoke version of GPT. Two, that really reflected my poetic style and my poetic concerns, and was much better than the off the shelf model at writing the kinds of things that I felt were valuable. So that's kind of the My main interest really is in using, using these tools in that fashion to kind of write with as a way of sort of thinking about how to expand my own practice as a writer. And in terms of thinking through what it means, you know, for the trajectory of literature, I'm someone who's been studying language, my entire life. So thinking kind of about what AI and what AI powered literature means on the grand spectrum from, you know, the oral tradition through the advent of written language and movable type. Now that we have generative text tools, in addition to lots of other technologies that are changing the way that we talk to each other and communicate, what does this really mean for how literature will look going forward, what it will do to us going forward, how it will kind of create narratives, and how that will shape the way that we as humans move through the world. So I'm kind of thinking about it from all those different perspectives. So AI, for me is a tool and a collaborator. It's also kind of a conceptual cone OMA almost, you know, to think about a lot of these larger questions and thinking about the relationship between language and technology, much more broadly than, than the mechanical level. I'll leave it at that for now.


Nitcha Tothong  8:19  

Max, do you have anything to? Like? What what is what about you? What is your day to day and like what you interact with, like creator with light? Ai practice?


Max Sills  8:32  

Sure. First of all, thanks so much for for having me. This has been an amazing two days onwards. So I just don't even know where to start. The I think I want to go to some of the copyright stuff that was talked about for a second. Specifically, the idea expression dichotomy. I don't think anyone brought up to get copyright protected fast, shouldn't also be able to cover it someone's ideas and mid journey that, that our goal, our mission, and the thing we were still trying to do is enable people to think and think in new ways and then communicate those thoughts to each other. So it is still very surreal for me to see like how much people are focused on selling the images or tricking each other with the images. I think the the negative uses of the technology get a lot of press. But what people don't talk about is and which which I see every day is the way people are exchanging images and creating new aesthetics and new ways of expression like 100 times a week. There have been so many like new aesthetic movements within Midrand someone needs to go in and archive all that incredible creative product. So I think the discussion is Wait is way behind. So one way I use it is it helps me think I Don't want to sell my images that helps me interrogate my own thoughts and feelings, I'm not very good at drawing. So I put in some prompts and I play with it. And I use it as a reflection tool. I'm also really interested in LLM. So for our, for my legal practice, I've started simulating contracts. And I, it's people people come in, like they asked for, they want a contract. But they don't ever actually want that. What they want is like, their entire relationship with the other person to be okay. And they want the whole thing figured out. And so I have these miscommunications, like over and over again, where people are like, can you please write this piece of paper, and then everything will be okay. But it won't be okay. If they don't trust the person they're doing business with won't be okay, if they're afraid. So, what I've been doing with LLM recently, and like, really, it's crazy for me, but you know, it's not like anything like you do, but I simulate for them, we kind of do print out stuff like the worst case scenarios, the best case scenarios. And I think that's going to happen a lot more frequently Thank you entire way people do business is going to change.


Nitcha Tothong  11:18  

Whenever you salvage from your experience, like what is like to be to interact with, like creator with AI, or like yourself with AI.


Sarvesh Mahajan  11:30  

So I have to start with the standard lawyer opening line, anyone know what that's called? Disclaimer. And the disclaimer, usually is anything I'm about to say is going to be completely useless. So do keep that in mind. But in this context, I think it is not that it's useless. But also an acknowledgment that we are at the beginning stages of something, which I think most people will say is quite revolutionary. I think it's going to have a lot of societal impacts, whether you look at it from intellectual property perspective, from you know, there's obviously a lot of worry about jobs and those kinds of things. So you know, we're, we're learning and so from that perspective, I don't, I don't feel bad saying that you know what I'm about to say it's useless. I'm here. Even though I'm sitting on this chair on this Dyess, I'm here as much as student to learn from people like Max and Sasha, and each and what they're doing and see what we can do with it. There was one thing that Max mention, which I found was interesting, and will also be my answer to the question of how do I use AI? And the answer is at least as a professional, I don't. And part of the reason is twofold. One is that's the guidance from our general counsel. And and the reason is because a lot of the input information would actually reveal client confidential information. So do I actually implement an AI tool in my day to day practice? No, but I'm curious to know what it does. And one of the curiosities that I've touched on is, contracts, which Max mentioned. So I'm going to throw a question out to the audience, if I may. How many of you have used chat GBT? I imagine pretty much everyone, right? So do you all know that you actually have a contract with open AI? You've agreed to set of terms which they consider an a binding agreement, and likely most courts would agree with them as well, that the types of information that you're revealing and so forth, and the your uses of their output are pursuant to the terms that they've that they've dictated. I looked at MIT Germany's Terms of Service, and I said, I'm not going there. Very well read. But yeah, I think the reason I raised this point is we've spent the panels that I had an opportunity to hear are very much focused on copyright aspects and other legal aspects. There is an underpinning which have contract and that also is something that is worth considering. And one just final thought on contracts is that I basically that's what I do I rewrite negotiate contracts. Sometimes I write some pretty mean Terms of Service to and contracts are really based on bargaining power. It doesn't mean that they're right. But when you enter into a relationship, what you're looking at is Can I can I, you know if if I want to negotiate He doesn't know if Max would even allow me right, he would just say, Look, you either want to use mid journey service as it's defined and as the terms are, or you can leave. So I know that I don't really have a lot of bargaining power there. In other cases, you may. But in any case, what I'm trying to get to is that as we contemplate gender generated by AI, particularly, but all the predecessor and sort of technology that's leading up to it, a lot of it is going to be dictated based on who has the bargaining power. And I think that's something that we should, we should be thinking about in these conversations as well.


Nitcha Tothong  15:36  

Really interesting. That also touched on something that I would love to talk about. But also like, when I to come back to such how what you were just mentioning, before also, Max, you touched on it a little bit on like, how the AI to inform your practice or like inform what you're thinking like how that process of like feedback loops like that, enhance or shift the way you're thinking, and what is missing from that process?


Sasha Stiles  16:16  

Well, I thought something that Max said, really resonated with me. And it's something that I've been thinking a lot about, too, which is that, when we're using these tools, it's not just to create something specific, it's to actually think through and kind of unlock or articulate, articulate something that we might not otherwise be able to say or see, or to draw or to paint. And even though these tools are called generative tools, it's not just about you know, generating an output. So, you know, I've been over the past, you know, two days been hearing the word interrogate a lot. And that is another word that I'm finding incredibly interesting, as part of this discussion, I use these tools quite a lot to probe and to interrogate, and to sort of help kind of turbocharge my own research, or to sort of speed read through things and kind of make links and associations that, you know, poets and writers have been making for time immemorial. But I'm able to do it. Now. It's speed and scale that, you know, wasn't possible for someone like TS Eliot, for example. So, you know, I think that's a big part of it, too, is that these tools are not just about making, they're also very well said, like they are about thinking, and they're about sort of enabling us to make connections between data points, in a way that we as individual humans, with our analog, you know, minds are simply not equipped to do. And I think, for me, being able to approach AI that way, is really revelatory, and dials down a lot of the fear factor, you know, in some ways, of course, but I tend to think about it really, as you know, there's the term artificial intelligence, I know is kind of a strange and troubling term. And I tend to think about it as augmented imagination, or amalgamated imagination. And for me, it's really a way of sort of combining the questions and the curiosity and the wonder the ideas that I already have, and then kind of strapping on a prosthetic imagination, and being able to sort of rifle through, you know, a histories worth of information that's right at my fingertips, being able to dip into the entire library of Alexandria without having to have actually read any single one of those books. And I think, you know, again, for me, as a poet, it's made me think quite a lot about all the ways in which I'm consciously quoting, or referring or alluding to other creative sources, and then all the things that are just kind of innately part of the cultural canon that I've absorbed, you know, just as a human being that I've learned over time. And so thinking a lot about, you know, my role as a writer, as someone who's supposed to be originating language, thinking about how that fits into this larger question, I think has, has been has been a very sort of revelatory way for me to approach this. So yeah, I really appreciate it that way of framing it.


Max Sills  19:09  

Yeah, I think like you said yesterday, language is a tool. Language is a technology. I, it's crazy. The difference. Sometimes I just feel like my thoughts are everywhere. I can't focus. When you open a notebook and you start writing. And then all of a sudden, I didn't know I felt like that. I didn't know that my thoughts were organized. And I, there's definitely this extra. There's that extra power of ideation that you can do. One, one thing that's that I've been thinking a lot is I see all these artists on MIT journey, making new things. And again, like the contrast between that and I think, caricature of appropriating existing work and then just trying to duplicate it, which I don't know where this idea comes from, because it's not It's not what people are actually doing. I think we're probably projecting our fears on it. But I think a positive of that is that the same old human evils are at work here. So you don't have to be afraid of AI just be afraid of human being like expropriating other people's cultural heritage like that, that's a human, this free bargaining power and using that unethical weight that's human. This is it's all just human behavior, maybe it can happen a little faster. But really, AI is more of like, a blanket that people used to cover themselves. Like, it's not going to materially change people acting ethically.


Nitcha Tothong  20:47  

So like, with that collaborative process, and the pattern that emerge, do you think like, these two, how these two can do better in terms of, you know, like, I Sorry, I'm gonna step back a little bit. It's really interesting when you talking like about, like, the terminology, like, machine learning artificial intelligence, cuz like, cuz like, to me, like, artificial intelligence is kind of like this tense, and make it intelligent is kind of like Alien in a way. But machine learning is really straightforward. And like, what if, like, what the algorithm like really do, like, your training you making model and then training like it learning from from you. So it has like that collaborative aspect. And, like, what you touched on it, like definitely, like, there's a lot of like, the system that we already have in hand like the boundary of like, the languages, or like the boundary of like, the power structure that we already have that definitely like adapt on these like AI tools, like, so I was wondering, in terms of like, how, how we do better in terms of like, she, like made the shift in a way that is like, it's a collaborative effort to make a better, better create the world or like better world? Yeah.


Sarvesh Mahajan  22:32  

All right. Like I said, anything I say is not going to be useful. But I'll give the question a shot, maybe just to get some discussion going. I think it would be it, the idea would be to address some of the evils that have seeped into society. So if AI is going to, you know, be used or become a prevalent technology in the world that we live in, is are there ways mechanisms through laws, contracts through other through norms, other ways of maybe balancing the bargaining power to use that example, reduce, maybe reduce exploitation to, you know, not, not to have aI sort of as a replacement for human human labor and effort and income, but to, to augment it and enhance it? So I think that's sort of I'm saying this not as a lawyer, or anything that I practice, but just some of the things, you know, when I'm sitting sleepless at night and wondering if there's gonna be a generator that can, you know, write all my contracts for me, you know, these are the things that I think about, we're all looking for, we've always I mean, I think it's sort of the perpetual effort or endeavor of humankind, to find better ways to live and do things. So it's just a hope that these technologies we have will, will help us do that. Can I? Please like, you can't pick on me


Max Sills  24:19  

if you go back, so because we still work that we think about what we do. But I still don't know what the answer is. I was hoping that we could talk about it together, you could help me think of the answer. I still don't know what kind of society we want to see, in terms of what system of laws or incentives maximizes creativity. I think if you look at our Constitution, that's the whole reason we have all these IP laws, supposed to advance science, it's supposed to advance creativity. That's why the government lets people enclose their own work, not because it's theirs. They did it. And their effort like inherently entitles them to that. Because the idea is, if you let someone say that something's theirs for a limited period of time, theoretically, that's close to incentivize them to be really great. But I've been thinking through this, I don't know what the answer is. But if we all agreed that we want to see a society where the maximum amount of people are being created. I know my brand stops there. Like, what what does the commons look like in that? In that world?


Sasha Stiles  25:35  

Yeah, well, okay, so I guess jumping off that, and taking it maybe in a slightly different direction. But I mean, it's something that I've been thinking about a lot, just in general. And then of course, today is just the scale and scope of the challenges and just this task at hand, when we think about the underlying models that are at the heart of all these discussions that we've been having, they are staggeringly large. And the idea of, of creating those and kind of retro actively making them good and better, is almost an insurmountable challenge in some way. And I guess like when I think about, and I relate this to like, the work that I've been doing with, you know, with my writing, as I mentioned before, like I'm working in a way that's very collaborative, where I'm actually fine tuning and creating bespoke and custom generators and things like that, to me, I'm sort of thinking about the fact that the way that we're thinking about AI right now is as this massive, massive, almost like too big to understand entity. And for me, what's really been helpful is kind of letting those models in a way be the stuff that is everywhere out there, that like is just around us all the time. And then to make it useful for me as a creative person. But to make it useful for anyone, I think maybe there needs to be another level on top of that, which is, you know, your personal point of view, your style, your information, your references, your personal history, your experiences. And putting those two things together, is a way of sort of saying, yes, all the information that's out there in the world is full of very troubling things. And that's always going to be the case. And then you know, that is that is there it is shaping the world the way that we know it. And then there is this ability to, you know, to use technology to, to create something to create a version of that, that's more individual and intimate. And that enables you to kind of engage with these tools on a very different level. And to maybe concretize that a little bit, it's like, I think about, you know, when when the computer was invented, the computers that existed were these massive room size pieces of hardware, that you had to go, you know, use at an institution or something, no one had access to these tools on their own. And then the rise of personal computing came and everybody had access to these tools. And now, you know, we can see how that's changed everything. So to me, I'm thinking, you know, maybe one piece of this, that is valuable to think more about is how do we create more of a personal AI? How do we create small data sets? How do we focus on you know, taking, taking what's large and using the best of it, but then applying it with something that is more meaningful to us on an individual basis. So I think personal AI to me is like something that feels like it is a valuable way of potentially like moving the conversation in a slightly more fruitful direction.


Nitcha Tothong  28:27  

Really interesting. And I wondering flash ha like on that note, like when you're working with like, their personal data? Is there anything that you feel like if cannot be contained in those, like, the tool that how it shapes? For example, I'm speaking in also Thai. And many of the modal and many terminology is like in English. With the translation, there's less shift in the meaning and shift in what it contain actually. So like with with that idea, like, is there anything that left out? That, that if you were not like using AI? And was this like the version of like your AI Alter Ego, let's say? Yeah,


Sasha Stiles  29:20  

yeah, no, that's a really good question. Well, so just to I guess, to like, put that to embed that in a in a concrete well, to concrete example. So you mentioned before that I've been working with bina 48 for a while now. So bina 48 Is this this human humanoid AI powered robot built by Hanson robotics was built I think around 2010 and is a cousin of Sophia robot who's like the much more famous one to come out of Hanson. But But bina 48 is really interesting because this whole project is basically an experiment in digital immortality and and how you can use information about yourself to sort of create a repository have data that will enable someone in the future, say, to interact with this mind file as though it was you and learn something about you. And so it's basically a question of saying, how, how much? And how much about yourself? Do you need to sort of download? And in what format? Do you do that? And then how do you train it? And how do you sort of shape it so that it becomes sort of a useful representation of you. And so that being able to work on that project has made me realize that you know, the gap between what's readily available, of course, out there, on the internet, and all that, and then your own personal history. And of course, there's all the things that we put out there in terms of social media and the things that we're sharing, when we post and all these different places, or read these terms of service and whatnot. But then there's also lots of other things that are sort of just, you know, that we take them for granted, or they're embedded in our physicality, and they don't actually translate to data in some way. Yeah. So thinking through a lot of those things has been, has been very top of mind for me and thinking through like all the sort of haptic information that doesn't get captured in, for example, in MLMs, or, you know, other other underlying technologies. And that's something that I'm very keen to think more about as a poet is how much of you know what, what shapes the experience is something like poetry, for example, is not just the text itself, but how much of it is also, you know, spoken word, or how much of it is aesthetic and seeing something on a page in a certain way? You know, and that relates to, you know, potentially the rise of things like, you know, from from text to image models, now, looking at things like facial recognition to image and like looking at how a model may be able to scan your face and catch an emotion that you don't even maybe know is there and turn that into an image. So thinking about? Yeah, I guess the language lists part of information that might be useful to get at some of some of these areas that are missing. But it really quickly. The other thing that's more of a concrete example is that one of the reasons I'm very interested in, in small data and in building my own data sets is because there's certain parts of my own family history that I know are just not really readily available anywhere. My mother is from a tribe of Mongolians called the comics. And it's a very small tribe, there's only like 3000 Comics in the US. And it's a dying language that is not really represented very much online, or in any, you know, archival repositories, there's a lot that's oral tradition, that has not been, you know, concretized anywhere and not inscribed anywhere. So I'm also very interested to think about how history is like that can be downloaded and preserved and the right way to do that, so that they're accurately representing the orality or the physicality of the tradition. So yeah. Yes, great question.


Nitcha Tothong  32:57  

So I think we are at time, but I wanted to ask if anyone have a question, who wants to open a question from the audience?


Speaker 6  33:11  

Hi, I, my question is, do you think there's a point in which the cost of not using generative AI will outweigh the legal or moral precedent? So like, maybe this is an extreme hypothetical, but say a law firm decides to allow the use of cash EBT in their line of work, and as a result, benefits from that expediency and efficiency? And that sets the precedents for the rest of the law firms who can keep up. So in your respective bodies of work, do you ever think that point could exist?


Sarvesh Mahajan  33:48  

I think we're probably already at that point, the demand, at least as a lawyer to be more efficient has always been there, and particularly, maybe known to many of you that law firms traditionally have valued their advice based on the amount of time required to produce it. And now, all these technologies are really shrinking that amount of time. So I think we're there, at least in the private sector, we're all trying to figure out how do we do it as quickly as possible, while maintaining the integrity of our ethical obligations to protect client confidentiality? I don't know if that's a complete or direct answer to your question, but I do think that any profession that needs to produce sort of written content or you know is always under pressure to do it in a more efficient way. So other types of professional services accounting work, consulting work there's there's a lot of pressure right now to figure out how to do it but how to do it correctly. No, I don't I don't think unless look, unless bar comes out and says that you know what, forget this attorney client privilege stuff. But no, we're we're not going to, you know, put ourselves at risk for that. But I think the demand is definitely there. And we're maybe even stalling to meet that demand by using the the excuse of, well, I can share my clients information with this tool.


Speaker 7  35:25  

Would that not be solvent by borrowing the practice that artists are using with their own data sets? Play more on some of the other. Even, you know, building your own your own AI, you know, we all have massive data sets, right? We all have massive stories on our phones and digital archives. We're all archivists. And none of those different ethical concerns or contractual concerns exist, right, if you're using their own companies hard drive to, you know, make more efficient, the contractual agreements based on their own past contractual agreements that you're not signing, not something. It's really large company ones, that you're the only ones that are where we need to deal with some of these ethical issues. Because you're, you're using your own data set. And we know exactly what that consent are.


Sarvesh Mahajan  36:27  

Look, I think, open AI just released a version of GPT for business where the, you know, there's going to be different terms that they're going to govern, you know, how that's used differently. Maybe other companies are thinking of the same thing. So there's multiple ways of dealing with that. The original question was, have we reached a tipping point, and I think we have a lawyer may have views as well or any.


Speaker 8  37:13  

First, I think he has all for the panel, you know, especially to kind of riff on what you talked about in terms of, you know, your heritage, and you know, the language and thinking about how AI not only can be used to create language, but also think about how AI can be used to preserve language, right. And if we think about the power dynamics of the race to who has the best algorithm, right, and then we're seeing, you know, Microsoft entering the fray with their investment of open AI, right, we're seeing Google create one, meadows creating one mid journey, we have all these companies that are racing to it, and it becomes a power dynamic of who has the most money to make the resources they have the resources to be able to generate, how do we worry or think about preservation? Of like, the common the heritage is right, with like language as one example? Like, who? How do we think about the right methods by which we go about doing those preservations? Like, who's gonna be in charge of it? Is it people at meadows that people at mid journey, or are they like hiring people like


Max Sills  38:19  

that? Yeah, I can? That's a great question. I think one cool thing about generative AI is, I think, in a lot of ways, business and economics have become dramatically simplified. So it's just about chips now. Are there enough GPUs to do all this stuff. And so it could be as simple as the government, or nonprofits just get an allocation or somehow secure an allocation of chips to do inference? I really think it's important to not depend on big companies to do beneficial things for the community. It's, it's great when they do. But I think ethics washing is a it's a huge concern. And if people want to take I think it's probably the responsibility of those communities in the government to take charge. I don't think people should.


Sasha Stiles  39:19  

Yeah, no, I mean, I agree with that. I would, I would say also and and add to that, that something that I think is really interesting, that doesn't, maybe you take precedent in the conversation as often as it should, like, people kind of think about, you know, scraping data, and then like, everything is about that data and like how companies are managing, managing it or not. And there's actually a lot of creation of new data that's happening, like specifically to go into these systems, and they're actually, you know, okay, I'll give you again, you know, my experience as a poet is that usually there's not very many ways to make a living as a poet that actually a lot of these companies are starting to hire poets and short story writers and MX and, you know, other other writers who normally don't really have that kind of a livelihood, they're, they're hiring these people to come in and like write, you know, write creative samples, you know, write all sorts of texts that can, that they can then bring in without a lot of the legal, the legalities and the, you know, the concerns we've been talking about today. So they're creating original texts as samples. And so, you know, maybe there's opportunities in thinking about those kinds of new roles that are being created for these kinds of transmissions to happen. And, I mean, I'm very realistic about these technologies in general, I don't, I try not to be like pollyannish about them. But I also realize that we kind of create what we envision and what we think about. So I try to be very optimistic that a lot of the tech companies kind of like, realize all the scrutiny that they're under. And there are a lot of good people working, you know, at the level of hiring these people to come in and create data or hiring artists like me to come in and sort of use the tools and like respond and give feedback and all that. So I also kind of, you know, I look at my sort of strange experience of being an interloper in the tech world as being a perfect example of how things like this can start to at least become a conversation, bringing perspectives that wouldn't normally be at those tables. It's really important and that's another reason why I as a poet, and like very fed very strange to me sometimes on panels with lawyers and with tech companies, but I'm doing it because I think it's important, especially for tools that involve language at such a deeply embedded level. It's important for people that use language like all sorts of writers to engage with them and give feedback and you know, imagining ideate ways for them to become better and we're useful tools


Nitcha Tothong  41:53  

thank you for the love a perspective and hopefully we move forward with like these AI with care. So thank you so much.


Announcer  42:10  

The engelberg center live podcast is a production of the engelberg center on innovation Law and Policy at NYU Law is released under a Creative Commons Attribution 4.0 International license. Our theme music is by Jessica Batke and is licensed under a Creative Commons Attribution 4.0 International license