Code, Clones & Creativity w/ Taryn Southern
EPISODE #45
Apple Podcasts | Spotify | Google Podcasts
Artist Taryn Southern shares her insights into how collaborating with robots can unlock new creative possibilities, the experimental processes behind developing your own artificially intelligent clone, and the tools we can use to preserve our digital legacy and animate our ancestors.
Taryn Southern is an award-winning artist, storyteller and strategist whose work explores the intersection of emerging technology and human potential. Some of her projects include: I AM HUMAN, a 2019-Tribeca Film Festival documentary exploring the future of the brain, the world’s first AI-composed pop album to hit the radio charts, an award-winning Google VR series, and the world’s first Ethereum-enabled song token. From biotech to blockchain, Taryn’s work has been featured in publications like Vanity Fair, Fast Company, Wired, TechCrunch, Business Insider, Forbes and more.
YouTube
SoundCloud
Transcript
Luke Robert Mason: You're listening to the FUTURES Podcast with me, Luke Robert Mason.
On this episode I speak to artist and futurist, Taryn Southern.
"I just get really excited about the future of historical preservation, and story, and narrative preservation, and heritage preservation - through these different technologies like AI that will enable us to repackage our experience." - Taryn Southern, excerpt from interview.
Taryn shared her insights into how collaborating with robots can unlock new creative possibilities, the experimental processes behind developing your own artificially intelligent clone, and the tools we can use to preserve our digital legacy and animate our ancestors.
Taryn, I became aware of your work when you created a music album in collaboration with artificial intelligence called 'I M AI'. So what was it that you actually learnt from that experience?
Taryn Southern: I learned a lot, you know. It felt like the early days of AI in many ways. There were a number of applications out there, but they weren't consumer friendly. They weren't user-friendly. You had to know how to code in order to create music with AI. So the barriers to entry were quite high, and I was fortunate. I reached out to a number of companies that were early in the space: IBM Watson; Amper; AIVA; Jukedeck. And I just said, "Hey, I really want to experiment with your technology. I only know basic coding. Can you help me?" All of them said, "Yes, this is what we want. We want artists to use our product. We're basically just building for a bunch of engineers right now." So for a number of these companies, I was sort of the test case for their product or software and was able to be intricately involved in the process.
I took home a few things. One, that creativity doesn't have to be so narrowly defined as a certain type of process. A lot of people will say, "Composing music or making music involves X, Y, Z. It's sitting down at a piano, writing the chords and coming up with something beautiful." That's certainly one way to make music, but there's this whole other process that I suppose I refined in working with these different technologies, that required a slightly different set of skills. Because I had just directed - or was directing, rather - a documentary during this time, I was able to easily sort of see the similarities between directing a film and essentially directing artificial intelligence to make music in the creative vision that I had. That's its own skill set, so I very much likened it to that process. That was one of the things I learned.
Mason: It's interesting that you say you directed the AI. The thing that I always wondered is: why was the music not released as 'The AI featuring Taryn Southern'? Why was it important for you to still have some form of ownership over that work, despite the amount of work that was done by the amalgam of artificial intelligent software that was available to you?
Southern: It's interesting. We ended up getting into all of these philosophical and pragmatic conversations around, "What is composing music? Who is the composer?" Think about it this way. If you had heard any of the early outputs of the AI from my album, they would have been almost unrecognisable. Think about it like this: As a director of a documentary, I'm shooting hundreds of hours of footage - sometimes thousands of hours of footage. I've got to take that footage and make sense out of it. I've got to create a story - a through-line - and there's a million different ways that you can go with that raw footage. If you give that to 10 different directors, you're going to have 10 completely different films and styles. So ultimately, I'm taking a ton of output. I'm giving the AI direction. I'm saying, "I want this and this style with this BPM, give it to me." Then I'm getting 99 pieces of turd music and maybe one piece - or a few pieces - of good music. Then I'm cobbling it together. I'm editing it together. I'm sort of synthesizing all of it into something that I feel is interesting and beautiful. The end result really does have a huge amount of human input, and thus does feel more collaborative. But you could argue that the actual raw material is all the AI.
Mason: Well, that's the interesting thing - what you said there about turd music -...
Southern: [laughing] I know. I threw in a very sophisticated term there.
Mason: Well no, we can make it an official term for the detritus that AI algorithms can sometimes output. But I guess in a funny sort of way, that work is what's more interesting to me. The fact that this was advertised and the press picked it up as this incredible collaboration between AI and humans, but in actual fact the AI needed a lot of coaxing. There was so much output that was not fit for human ears. The question then becomes: If it wasn't a fit for human ears, why did the AI see it as this potentially beautiful, valid form of output?
Southern: Well, because the AI doesn't have a value judgment - yet - around the quality of the music.
Mason: [laughing] Yet!
Southern: But we will get there and it will probably be trained on neural algorithms taken from humans that are listening to these pieces of music. Then they're getting real time reactions, and then the AI will learn from that. I mean, there's no doubt about it that we'll get to a point - and probably sooner than later - where what I did on that album is no longer necessary. That becomes a whole other philosophical discussion, but at least for the early days of AI, for really any kind of art - painting, digital art, music - there's a lot of handholding needed with the human. Then you could argue that it's also extending human capabilities in a lot of ways. I grew up playing piano. I grew up listening to Spice Girls and pop music. You could argue that my neural algorithm is trained on a very specific set of musical influences. My knowledge base and my ability to create is going to be somewhat self-contained in a space based upon what I have learned, so me collaborating with an AI that has all of this knowledge and breadth of expertise in jazz music, in 1600's classical - in all of these different forms of music - is going to expand my creative repertoire, and it did. That was really the other second takeaway that I learned from it. That's the exciting part for humans that are endeavoring into anything creative with AI.
Mason: So in that case, why not limit the output of the AI? Why not create a system that would just listen to Spice Girls music, to output something that you would find - or certainly nineties Taryn would find - extremely appealing?
Southern: You can. You can train them on those specific sets, and the more you train, the better they get. It's still a learning curve because it also comes down to who the engineers are who are building these algorithms. If you peel back the curtain - or peer under the hood, rather - at the code, how are they taking these learnings and then building off of it? That's really where there's also a lot of highly creative work going on, on the engineering side, too.
I mean, you're right Luke. These algorithms are going to get really good. I'm already hearing a huge difference now, which is three years after I released the album. It's incredible what these technology companies have developed in that short span of time.
Mason: Then why is the aim to represent music as it pre-exists in life? That's always been the case. Especially with EDM when that first came around - and before that when we had these tapes that were able to create these loops and all these different ways the BBC Radiophonic Workshop used to create music. With these new forms of technology, we had new sounds. Yet it feels like what these AI systems are doing is ingesting pre-existing human created music in the attempt to re-express human creativity. Why not allow them to express AI creativity? Even if AI creativity - the sound - is, as you expressed it, turd music.
Southern: Well, I know that there are a few artists working with AI music that absolutely love the turd music that's being outputted because they do find inspiration in the oddities and in the bizarre kind of textures that are being created. There is a lot of novelty happening, even just on the creation of sounds. I could use Google's...I'm struggling to remember, it's been so long. I think it's called 'NSynth', their AI tool, which allows you to blend different sounds. So I could take the moo of a cow, interject it with a synthesizer, and then add the stylistic preference of a jazz horn. Then I get this whole new sound, and it's really wild.
Why are people creating AI that's sort of building off of the human ear? I mean, there are a lot of different reasons, but I think the primary two - and I'm not saying whether they're good or bad - but I think one is this need to have cheaper, more readily available music for productions. The rise of digital media has meant everyone wants to do things cheaper and faster, and so AI music presents a potential solution for that. Then the second, I think, is customisation. Knowing that we're five years away, potentially, from everyone having some version of a brain interface in their home and wanting a customised musical playlist that's fit exactly for their brain activity and what their attention is that day. Do you want to focus? Do you want to find hope? Do you want to just cry your eyes out? Do you need to sleep? You run all that into the algorithm and the AI then builds you your own customised playlist.
Mason: That's a fascinating idea, but driving towards efficiency as you just said there - the ability to create music quickly and cheaply - surely that's the enemy of creativity? Creativity is about playing around, experimenting, and getting things wrong. If AI's aim is to output something that is correct, that is right, that sounds good - does that ruin the process of creativity?
Southern: Yeah, this is my favorite juicy question. I always like to play devil's advocate on this question, maybe because I feel that when most people enter the conversation, that's where they go. They're like, "This isn't creative if something is just sort of doing it for you." There's an argument for that. Yes, that you’re reducing the creativity around the music creation process, but if we look at any tool that has come onto the scene - that has made creativity easier and that has reduced the process to something more streamlined - it's hard for us to argue that it's inherently reduced creativity because those tools inevitably open up new possibilities.
Think about an iPhone, for instance, and how much easier it's made taking photographs. You used to, as a photographer if you wanted to take good photos, you sure as hell better have been educated and learned about f-stop and learned about exposure and all of these different things like depth of field, so that you can take a beautiful photo. Now you've got an iPhone automatically doing all of those things for you. Is it reducing the creativity of the photographer? Perhaps, but it's also opened up all of these other possibilities in VFX and video creation. Now, any 10 year old can essentially make a beautiful video with CGI graphics. So maybe we reduce creativity in one area, but we've sort of opened it up in another.
I think you could argue that AI music may very well be that. Once we have these tools that make things more efficient, where do humans then spend the extra time that they have with these projects? In my case with my album, my album did take a lot less time to make than if I'd made a traditional album. I spent that time making 360 degree VR music videos for these songs and learned how to create in VR. That was a really cool process. I really thought by the end of this album that with the future of artistry and creativity, people will look back on how we used to say, "I'm a painter, I'm a filmmaker, I'm an actor, I'm a singer, I'm an editor." And they'll be like, "That's so crazy." Either you're a creative artist, and you're doing all the things, and you have this vision and you're just bringing it to life through these different mediums. I think that is the future of creativity. People that are going to be able to play in all these different mediums and have this breadth. They'll be able to do that because our tools will be so good.
Mason: You can tell that you've come from a YouTube background where you have to be all of those things. You have to be the talent. You have to be the person editing it, the person distributing it, the person marketing it. You have to play these, these multitude of roles to be able to produce and allow your creativity to be seen. But should the aim always be to allow this stuff to find an audience? Thinking back to AI music, what about music that the human ear can't even hear because it's at frequencies and resonances that machines can create and pick up on, but maybe human beings can't? Why should creativity always be boxed into an acknowledgement of the fact that there is a potential audience for that creativity?
Southern: I love that question, Luke. If you could somehow solve that for us. I think that's more a reflection of where we are as a society than it is of these new tools. You look around and 13 to 16 year olds don't don't want to be astronauts. They want to be TikTok stars. Of course, I'm making a generalisation. I'm sure that there are some out there that would be excited to be an astronaut, but I can't believe how quickly we moved into a space of 'produce produce produce'. Whether it's monetary or a desire for fame, it's appealing to the lowest common human denominators - these technologies. So it's sort of pushing everyone towards that.
I don't like it. That is why I left YouTube. I left YouTube for that very question. I got into this space to be a renegade and to do my own thing, and to buck the traditional Hollywood system. That's how it started, but then it became something else entirely. It became this hamster wheel that you had to keep up with. If you didn't keep up with it, you would be punished. Not just by your audience but by the algorithms, because they were built for frequency of output. Unless we solve - I mean, there's a number of different things we have to solve - but until we do that, creativity will absolutely be commoditised.
Mason: Firstly, I'm terrified that the next generation will be twerking on Mars. That's if the TikTok generation finally makes it to the stars. But also, hearing you describe that experience of creating for YouTube, you were collaborating with algorithms. You had to play the algorithm to be able to get your work seen. So I guess this AI creativity, this collaboration - this is something we're doing by the nature of using technological tools such as our laptops and platforms like YouTube.
Southern: Yeah, absolutely.
Mason: The question is whether that's the thing that we should be focusing on?
Southern: What do you think that solution is, Luke? What is our starting point, even?
Mason: I'm fascinated by - even with doing this podcast - I'm fascinated by conversations that aren't captured for distribution. Even this right now, we're aware of an audience. This is going to be listened to. We're currently in the ears of other people, you know? So we have to then purport ourselves and present ourselves in a certain way to ensure that a conversation is flowing and to ensure that it's going to be interesting to a perceived audience. It's getting to the point now where every single conversation on planet Earth has to in some way, shape or form be captured for the purpose of distribution. Why can't we have a conversation without a perceived audience? I'd love to just hang with you Taryn, but in a weird sort of way, doing the FUTURES Podcast is a valid excuse for doing that and asking you these questions.
Can we get to a point where we just create for ourselves? I wonder if that's why you're creating these AI clones. There's a realisation now. "Hey, I just want to do my Taryn Southern stuff, and I'm going to let my AI clone go out into the world and play that mediated game on my behalf so that I can reclaim my brain, and reclaim my creativity, and not have to produce for algorithms." I'm fascinated by people who clone themselves. I think I should let our audience know that you have cloned yourself. You've created this AI Taryn. So who is AI Taryn? What is she? Or what is it, I guess? How does she, or it express itself on the web? Then I guess we can try and work out what that means for the future of these platforms.
Southern: Oh, goodness. Yes. I mean, I'll just preface by saying I'm still figuring it out.
Mason: Aren't we all?
Southern: We are. By the way, I just really want to double down on what you said there. I think the fact that we are all having so many performative conversations is hindering a lot of nuanced perspectives because when you know that there's an audience to be had, either people are going to be more measured in how they present their opinions or they'll take the megaphone shock approach and just shout the craziest thing to get attention. It's not actually allowing for a lot of nuance to come through. So I appreciate that perspective.
On the AI clone front - which is a different thing entirely - about a year ago, I was put in touch with this company, Hour One - they're based in Israel. They had been working on, essentially, a re-imagining of what AI video could do if you're recreating pixel by pixel a human image. Up until they came along, most of the AI video technology that we've seen was this kind of mesh technology. You're sort of meshing images and the algorithm is taking the average and then presenting that in pixel form - otherwise known as deep fake. You'd get these images of people where you could tell that there was something weird about them. They were uncanny, but something was a little strange.
So this company came along and said, "Hey, what happens if we have the algorithms learn through this different process and recreate the images pixel by pixel?" And that's how they began. The initial use case for the technology was education. Let's take teachers who are the best teachers in the world. They don't want to be on video, teaching everyday. They don't have the time for it. We need to scale out their abilities. They can't speak in different languages. So what happens if we clone them? They then submit their curriculums. Now we can have that translated in every single language, ship their incredible curriculums around the world, and they don't have to show up for class. They get to just focus on learning and teaching. That was the initial use case.
I was put in touch with them and we decided to go ahead and make a clone of myself with the idea being that this is a big experiment. What happens when a creative person - who is an artist in front of and behind the camera - has one area of production streamlined in a different way? What does that open up the possibilities for?
Mason: Well help our audience understand. You said it so matter of factly. "I've just cloned myself." as if it was something that you can just do every single day. "I've just cloned myself again." It's like, "Oh Jesus, Taryn, how many of you are there?" It goes back to that very Marshall McLuhan thing of human beings being fascinated by versions of themselves in other media. You've created a version of yourself. It is in another media, but it's still a representation of you. It's important that it is your face. Why make that sort of decision? Why couldn't AI Taryn have three eyes, two noses and six mouths? Why must it stay within these parameters of something that represents you as an individual?
Southern: Well, I think that I will soon have the option to add a third eye, or a horn on my head, or whatever I want. That's also very exciting because real Taryn can't just costume-change in 10 seconds. For me as a YouTuber - when I was doing that - the whole thing was about connecting to your audience. There is something about having your face on camera that matters. I think in doing this AI experiment, part of the experiment is: Do people still connect with this clone version of oneself? I have no preconceptions about whether they do or don't. I think that there are also a lot of challenges and limitations around the technology that might prevent someone from feeling as though there's an authentic connection.
That also opens up a whole other can of worms. Is it bad? Is this a Black Mirror episode, to have all of these people connecting to someone's clone? Part of my experimentation is figuring out the answers to those questions. What's interesting for me is that I have very little to no interest in being in front of the camera anymore. I did it for 10 years of my career. I had a ton of fun and then I was done. I was exhausted. I was worn out. I was also like, I'm not that funny. I'm trying too hard. Why am I pushing so hard at this thing that I'm mediocre at? I really, really love ideating, writing, directing - that lights me up. Yet I built something over 10 years that involved my face and likeness and personality. It's not that I think that that needs to be exploited further, but there's an opportunity for me as an artist to say, well, I know myself really well. If I take the best of all of my work over the course of 10 years, I can feed it into an algorithm that can actually make me a better performer, funnier, with better comedic timing - all of these things. Then I get to direct myself, without having the bias being on camera, or the nerves, or whatever it is that's going on that's preventing me from my best performance. I just get to direct my AI self. What a fun experiment to see what happens. Maybe I will make AI Taryn 20 times more entertaining than real Taryn ever was, but I still get all the credit. It sounds great.
Mason: Great. Well, I mean, that's an interesting point - that you still get all the credit. At what point does AI Taryn turn around and go, "Hey man, I'm making you funnier, smarter, and way more interesting. This is largely my work right now. Why am I not being reimbursed for it so that I can buy digital things and bits that I want on the web?" But it's interesting to me. I guess it kind of makes complete sense that someone who spent so much time connecting through a lens for YouTube created this version of themselves - this mediated version of themselves - because I guess Taryn Southern the YouTube sensation was always a mediated expression of who you actually were as a breathing, biological human being. There was a disconnect there. It's very Erving Goffman - 'The Presentation of Self in Everyday Life'. You were presenting yourself for your audience. So it kind of makes sense to just outsource that shit to an AI.
Southern: What I'm about to say would be, I think, blasphemous to a lot of the YouTube community. The reality is you can sit there all day and say to the camera, "I love you guys. You're my family. We're a community." No. You care about the people that tune in and you care collectively about what they think. There are real connections that you make, but most of the time you're staring at a camera lens. You are not actually having a real connection. That's part of the problem with all of this. I think people are confusing it with real connection. They're losing some of the basic, intrinsic parts of what makes us human and how we develop relationships and how we communicate, because young kids are learning how to do so through the lens of a phone, and it's just a different thing. I would not argue that that is an authentic connection. Maybe you're right, that in some ways the AI clone didn't feel as controversial or as shocking to me as it might to others. Cause I'm like, well, I've kind of already been doing this. I've had a fictional version of myself that was acting out for 10 years in front of a lens. Not feigning connection, but not really having a real connection, either, with the audience.
Mason: "Hey guys, like and subscribe"
Southern: "Love you so much."
Mason: It becomes its own language, you know? It's odd to also watch people trying to connect through these shiny glowing rectangles. I guess in a funny sort of way - and you've been quite open about this, so I don't feel bad talking about it - but YouTube led to a degree of depression for you. I mean, it was a real struggle to realise that this thing was giving you a large audience, but not necessarily giving you internal happiness. I think the realisation that you're creating character for these virtual environments is the thing that gives you back your mental health, in an odd sort of way. The sadness comes when these individuals curate these versions of themselves online. The sadness comes from the realisation that the version of themselves online is not creating the feedback loop back into their own reality.
I think the expression of that on Instagram is where you see these young individuals who are not just Photoshopping their body and their self, but photoshopping whole backgrounds. They're putting rainbows in the sky, pink tinting everything - quite literally rose tinting the world - putting fake stars or fake birds into the sky, and not just creating these new identities for themselves, but these whole new realities. Then they switch the thing off and have to go back to the grey, basic banality of their lives, where they're refreshing to ensure they've got more followers, or more subscribers, or more likes. In a funny sort of way virtual influencers could be the solution to the sadness that comes from trying to be an IRL influencer.
Southern: I mean it's quite possible. It's actually exhausting to put on a front all the time. Granted, I think that for today's influencers, there are a few major differences with the kind of environment that they're growing up in and the one that the OG YouTubers grew up in. I think ours was more performative. It was not okay, eight years ago, to talk about depression or anxiety, or to talk about what was really going on. So there, there might've been more performative aspects to what we were doing.
That being said, we also didn't have Instagram stories, Twitter or Snapchat. We started, really, with just YouTube, and maybe Twitter. So there wasn't this constant sharing of one's life in the way that there is now. I think for all the reasons you said, it's causing a lot of mental health issues. We'll see what happens over a period of time as these kids get older and they have learned validation. These are the social media, so what happens when that's taken away? Also, how does it change our brains? I mean these devices are powerful. These are strong hits of dopamine that are coming in every single time there's a post. Every single time there's a notification. There's an argument that it's actually changing the very makeup of the brain and how it's processing joy and different human emotions. That's really frightening as well.
Mason: Yeah, that's deeply concerning, but then that's also a good argument for entirely virtual influencers. In a funny sort of way, I've come to the realisation that virtual influencers feel more authentic than real life influencers. Because if it's true that the medium is the message, then these virtual influencers are quite literally a product of the medium in which they've been created. So there's lilmiquela - she's an entirely virtual Instagram influencer. She has no real life identity. She is the amalgam of the creators who've put her out into the world, and she can live authentically because that is the world she lives in. She doesn't live in the real world. So she's entirely authentic to the medium. I guess I trust them more. I trust them more than real life influencers because of that.
Southern: Yeah, I think that's really interesting. I haven't heard that perspective before, Luke, but I like it. It's something I'd have to chew on. I enjoy following lilmiquela, by the way, it's such a fun ride.
Mason: It's a fun ride. And it's something that you feel comfortable with because the desire isn't eventually to monetise these things. Perhaps it's not the platform that's the problem. Perhaps the problem is the business model. We're beginning to see more and more of these micropayment style models for creating content. It does make me worry that it starts to define, again, the form of creativity. So with Twitter launching micropayments where I can start paying you, Taryn, for your Tweets, and Instagram doing tip jars and God knows where else, and things like OnlyFans where it's basically about sending your nudes and getting payments for that - it's like every little expression of you has some form of value out into the world. The question becomes: As you start valuing parts of your identity based on what the market value for that thing is, does it ultimately cheapen who you are?
Southern: 100%. It commoditises all human behaviour. It also incentivises us to behave in ways that make us money, but aren't necessarily good for society - which is the biggest concern that I have. It transactionalises relationships as well. Is that a word: 'transactionalises'? Well, I just made it up, but I completely agree. I think that it's an interesting sort of conundrum that we're in right now. On one hand, this technology is moving at breakneck speeds and it's already hit a saturation point within a generation. It's not going away, and it is heavily influencing behaviour. On the flip side of that, we're seeing people really struggling with mental health but talking about it more than they ever have in the past. Certainly more than our parents' generation did. There's an awareness and there's a desire to learn more about what's going on in our inner world, and to make sense out of that. It's weird - as we're developing the tools that inevitably are leading us down one direction, we're also developing the tools that help make it easier to dive deeper into what's going on internally. I'm very curious to see where this all nets out and am hoping that I'll end up on the right side of history with all of this.
Mason: Well it's the argument for AI clones, really. I'd rather have a version of myself - Luke Robot Mason - that would go out into the world and do all my social media for me. I'd feed it my last decade worth of Tweets. It can Tweet on my behalf, and go and earn me some money. What I'll go and do is reclaim the real world. I'll go and sit in a coffee shop and read a book, or take a walk, or not have to interact in these environments. Yet the AI on my behalf will be making me the micropayments needed to live effectively as an influencer. It's a very compelling model, but when it comes to the business model of it, how are you thinking about licensing your avatar? Because it has your face. It has your voice. How do you start thinking about the legalities related to AI clones? How much of the thing do you feel you own and how much of the thing can you verifiably monetise?
Southern: Yep, it's a great question. It's still really early days, so there's not necessarily a market, or there's certainly not a rabid market for AI clones. I do have a deal with Hour One that allows me to essentially approve or disapprove anywhere my likeness is used. I think that there are two questions here. One is, what are the legal frameworks that are being developed around this technology and where do creators sit in that? These are real human beings, where there are clones being made of them. So there are a lot of protections that need to be in place to ensure that it's not misused. Secondarily to that, what is the economic marketplace? Is there money to be made that could allow real life Taryn to read a book, as you were saying, and play in the grass and whatever else I feel like doing.
It's still very early as I said, but my next goal in this is to help establish a sort of body of voices that can ensure that as more of these companies pop up and develop, there are protections in place. Almost like a union of sorts for on-camera talent, to ensure that this doesn't go in a really dark direction for them. Also to ensure that licensing payments are fair and reasonable. The thing that I do fear - and I'll just be open and honest about this - I do have a fear that the cost of these AI clones could become so cheap that it suddenly renders real humans obsolete. Why would a company hire a real life person if they can get the AI clone version who doesn't sleep and doesn't eat? It does everything they say perfectly at a fraction of the cost.
Mason: Well, does it? Let's flip that. Does it actually make the real Taryn more valuable? There's a digitised version of you or an AI clone version of you that can do whatever it wants, but then they could also have Taryn Southern - the real person - turn up and do a piece to the camera. Does that increase the value of the fact that it's authentically you that shows up to read whatever it is that the brand or the organisation or their client wants? Could the AI version of you just do the basic banalities, make you a couple of micropayments for reading Tweets, or doing the Cameo stuff? Let that thing do the Cameo stuff, make me a fiver for wishing someone a happy birthday, and off you go. The real Taryn can do the real work, the important work, the interesting, urgent work, the documentary work that you have been doing.
Southern: I love that. I think that that's certainly one way this could go, and I hope that that's what happens
Mason: Does it need to be Taryn? Could there be an army of individuals similar to you? This is the funny thing about virtual persons. Does there need to be a creator? Because when I see the new sort of wave of virtual persons that are based on these re-purposing of video to create something that represents a human being, I am always worried that there's no verifiable origin. Things like MetaHuman - Unreal Engine's version of creating purely virtual humans - are incredible. They're utterly beautiful as creatures. I'm going to call them creatures instead of humans. But the thing about them is that because there's no origin and because they're not based on anything, do they actually present themselves as demons? Are they embodying the worst of technology but in human form, rather than embodying the best of humans through technology?
Southern: Hm. Why do you say that? What makes you think that they could be embodying the worst of human form just because they're not based on something?
Mason: Because there's no responsibility. It's joyous when an artist creates an AI avatar version of themselves that they culture; that they curate; that they create - because it feels like they have some form of ownership and relationship to the thing. If you can press a button and a face appears that looks verifiably human but can do basically whatever you want, then I guess it has less value. It just shows that human beings are as easy as a button press. Whereas if you're working together with a form of technology, to train this thing to express its humanity, that version of an AI clone seems to have much more value than something that's an amalgam of a multitude of expressions of humans in technological form. I.e. when it's yours, it feels so much more vital and so much more important. Yet it then is almost like this owned - I don't want to say slave - but it's almost like a thing that you constantly have to check in with. It doesn't have its own autonomy.
There was an artist, Stelarc, who used to talk about third life. He'd say that second life's great. But second life are these avatar versions of yourself that you have to puppeteer. When you log off that thing doesn't go and live its own life in the virtual world. It doesn't walk around on your behalf, make connections on your behalf, make friends on your behalf. And then when you log back in, you check in with it and it goes, "Hey, I've made all of these new friends now, and you're going to like these people because I'm an expression of you, and because I'm an expression of you, I've made a friend with this person." These things don't make us better humans. They just are a confined expression of who we want to be in virtual form.
The interesting thing about AI Taryn is AI Taryn is a subjectively edited version of you. If you gave it pure and utter unfettered access to the multitude of expressions of you, if it was allowed to trawl through all your private messages and all of your interactions, if there was something ambiently following you throughout your life - whether it's perhaps a chip in the brain -
Southern: Terrifying.
Mason: Yeah. That's terrifying. The idea that this thing is not subjectively edited and the idea that you would give it complete and utter freedom to teach you something about you - that's where it becomes a challenge. As a mediated activity or as a form of creativity where you're crafting it subjectively, it feels safe. But if you just let that thing go at all your digital devices then boot up one day and talk back to you, that's the thing that scares people.
Southern: Yeah. Can we please not have it learn from 2010 Taryn?
Mason: Why not create a 2010 Taryn? Why not create an Instagram influencer based on 2010 Taryn who was doing all sorts of weird and wonderful things on YouTube that maybe 2021 Taryn would think are off-colour now? Why not let that exist as an identity out in the world and just relinquish it? Maybe it could be very cathartic as an experience.
Southern: It absolutely could. I'm so intrigued by this idea. I think it's fascinating to be able to see the evolution of self through the eyes of algorithms that are studying and assessing where the changes have been taking place. We also, as a society, would have to get way more forgiving. Oh my goodness. This would just wreak havoc.
Mason: Well, that's the thing. Your other interest is brain computer interfaces. These things that always remind me of Greg Egan story 'Axiomatic: Learning to be Me', where he talks about being six years old and having this small dark jewel inside of his head that slowly but surely learning to be me. Not to do a spoiler alert but long story short, that short story describes how his brain is scooped out and the jewel then replaces his brain, because that thing is basically learning how to be him. Why need a biological brain, when you could have a much more durable form of that? How can these things speak freely in a world of cancel-culture? Could it get to the point where we're starting to cancel AI bots?
Southern: Probably. I don't think we've seen the end of it yet, Luke, sadly. But something's going to have to change. Something is going to have to change because we simply cannot continue living in a society that has zero tolerance for ideas to evolve. We need to encourage idea evolution, and that means coming up with a different way of handling people's past transgressions, I think.
Mason: Well it comes with a new way of understanding identity to realise that these things don't have to have coherent narrative. When creating these AI clones or a YouTube video, or a documentary, you have to craft the narrative and it has to have a beginning, middle, and an end. I worry about two things - one that I know you're more involved with than the other - which is digital birth and digital death. How we subjectively edit a child's identity up until whenever the digital bar mitzvah is, and we hand over the passwords to the social media accounts that their parents have been subjectively creating and editing for them for the past 13 years of their life. We haven't had a generation that's lived through that, but the question is, what does that child do? Personally, thinking back, I'd just delete everything and start again. I'd be like, Jesus, why are these baby photos of me sucking my thumb on the internet? Then there's some blasted marketing algorithms looking at all these photos of me sucking my thumb, realising that I'm orally retentive and going, "Alright, okay. The best way to market to an orally retentive versus an anally retentive person is to do this, this and this." It's a horrible thing to be exposed in that way. Then with digital death, the question really is, Taryn, how do you want to be remembered?
Southern: It's such a good question. It's one that I started chewing on two years ago when I was diagnosed with breast cancer. It was an interesting situation to go from one week feeling like you're on top of the world and that you're as healthy and as happy as one could be, and then the next week you're getting a stage three cancer diagnosis and being asked to sign a will and medical directive. It brought up a lot of questions for me, particularly when I was going through chemo. I felt like I was dying even though I wasn't, but your brain is telling you otherwise. I started grappling with some of those questions.
Now it's been a year since I finished treatment and I've been consulting for a company called Good Trust, which basically approaches this problem or this conversation around what happens after we die. What happens to our digital identities, our digital stuff? Do we care about digital legacy? And if so, how do we have some control or some ownership over that while we're still alive? I thought about it. I don't have kids yet. So I'm sure that the answer to this question will change dramatically when I have kids, but there was a part of me that thought I would like to ensure that my friends and loved ones have some kind of interaction with me after I'm gone, whatever that looks like. Whether it's a collection of my memories, whether it's insights or takeaways that I feel might be helpful for them. Then when I think about kids and add that into the mix, I'm like, oh, I really want to make sure that they have some piece of me - hopefully the good piece - that stays behind. Then I think as an anthropologist - I got my degree in anthropology - I just get really excited about the future of historical preservation and story, and narrative preservation, and heritage preservation - through these different technologies like AI that will enable us to repackage our experience for others to learn from, and indulge in, far after we've passed. What I would give to have my great, great, great grandmother, who came to America on a ship, giving me advice from her perspective. How cool would that be?
Mason: Well that would be cool because your great great grandmother had this incredible life story. But there's a whole generation of kids who have grown up on TikTok, whereby their grandkids have got to see a hundred videos of them twerking on TikTok. It's like there is a question of what is valuable to preserve and to keep. There's two things going on when it comes to digital immortality. One is this subjective editing during your life of a version of yourself that your ancestors will find. On the other hand, there's this digital reanimation of the artifacts that we have, and that we've inherited from our ancestors - whether they're black and white photographs or diaries.
You look at someone like Ray Kurzweil. His drive to upload minds is really a narrative about bringing back his dead father. He has a storage unit somewhere in California, where he has his father's diaries. The guy wrote a heck of a lot of these diaries and his hope is that you can ingest that text and kick out something that is authentically his father. But it's not going to be authentically his father. It's going to be how his father subjectively edited his identity and how his father wanted to be remembered, because he was writing those things to potentially be read.
It does scare me a little bit that we are allowing so much of ourselves to be the raw material for the possible AI that will reanimate us. It could get it wrong, and we won't be around to say otherwise. When you see that the company you're working with at the moment reanimates those, they're beautiful. They reanimate those black and white photographs of your ancestors. The first thing I thought when I put my grandfather through that was: Did he actually move like that? Did he actually smile like that? He's got a weird cheeky smile. I don't think I ever saw my grandfather cheekily smile like that. Is that an AI's Interpretation of that photograph? So the first thing I did was take a photo of myself, put it through an aging algorithm so that it would age me up until about 80 years old, and then throw that thing through the software to see if it would animate the aged version of me in a way which I could recognise as me. Annoyingly it did, which I hate to admit.
Southern: Wow. I want to see the results of that, Luke. You'll have to send it to me.
Mason: I'll send it to you. Yeah, I would turn my head like that. That's slightly annoying. But back to the idea that there's two things going on here. Is there a joy in genuinely forgetting and in relinquishing your identity to the agency of other individuals? Because again, back to talking about AI Taryn, she is your creation and yet Taryn Southern lives in the minds of a multitude of individuals. You live in the mind of the 450,000 YouTube subscribers. They all have a different idea of who you are. I have a different idea of who you are in my head. Your fiancé has a different idea of who you are. Your grandmother has a different idea of who you are. You cannot control the multiplicity of identity. So why bother?
Southern: Oh, that's a really good and meaty question, Luke. Why bother? I mean, I suppose if you look at legacy, it's always been a presentation. It's always been a curated version of oneself, whether as a biography or something else. We choose to present certain aspects of ourselves that we feel are worthy of passing down to others. The interesting thing of course - that you've already pointed out - is that as time changes and culture changes, an AI may do a terrible job of presenting what we would actually want out there. But for some reason, we have this need as humans to want to preserve and share and pass knowledge down that we deem important. So I suppose I look at these tools as just another option set. People don't have to choose them, but they have them as an option. I just know for me, I would love to have my grandmother who passed away 10 years ago be digitised, and give me thoughtful pieces of advice. There is nothing that would bring me more joy and make me feel more connected to her and probably make me a better person than having this version of her in my cell phone that I could connect to immediately. Ultimately she's not here to tell me whether or not she's okay with that. And she's not here to watch, you know, the AI versions of herself back and give approval. We have to, at some point, I think, decide, are we okay with that? Are we okay with not having control over our future selves in exchange for future generations being able to have some sort of insights or information?
Mason: Do you mind me asking, Taryn, did you know your grandmother?
Southern: I did. I had a relationship with her until I was 23.
Mason: So you know that that individual is going to give reliable advice. But some people's ancestors might be arseholes. It's the worst thing in the world to have your ancestor following you around on your digital device, giving you advice when they probably gave very bad advice. You had a wonderful relationship with your grandmother so it makes sense that you want that person preserved.
Southern: Do you think maybe some things shouldn't be preserved? Is that what you're saying?
Mason: Yes, 100%, but I do understand the desire. The most compelling thing I ever heard was that we die twice. Once when the biological body stops working and second when our name is mentioned for the last time. I understand the desire to ensure that our name is acknowledged, recognised, spoken out into collective consciousness - or whatever consciousnesses. I understand that there is a need for that, but if we're not going to be around, why not allow that individual to navigate the world without this individual on their shoulder? Why create these pseudo deities that just kind of look over and judge everything that you do in life?
Southern: Sure, sure. Totally valid framework. I also wonder about preserving these entities even if these entities are not, are not good - even if there are real acts of evil that have been committed across history. You have to wonder, how did that happen? Like how did people allow that to happen? And there might be something in the preservation of that, that actually allows us to get to the bottom of it. I don't know, I'm playing devil's advocate.
Mason: No one is going to argue for a digitally reviving Hitler so we can work out why he was such an arsehole. No one's going to make that argument. It's the same sort of thought experiment, would you kill baby Hitler? It's like, well, baby Hitler wasn't a problem.
Southern: I don't know, but maybe it would allow us to study sociopathy or some other sort of pathology in a way that we really haven't been able to, and get better at sort of acknowledging our shortcomings as a society in seeing things that are heading down a really dark path.
Mason: The question is why do we have to do that with AI? Why can't we just do that with other human beings? Why can't we discover who we are not by talking to digital versions of ourselves or digital re-animations of our ancestors, but by just talking and interacting with each other?
Southern: It's not one or the other, Luke. I’m going for both, I'm going to vote. I'm going to be the moderation voice here and say that I think we need both, and that there's value in both of those things. But of course, we cannot lose the human to human connection. That would just be the biggest travesty.
Mason: Yeah. Ultimately, how do you think people are finding the AI Taryn experiment? Do they find it creepy or do they find it intriguing?
Southern: So one of my favorite things about trying these experimental technologies is having no clue how people are going to respond and knowing that I'm really taking a big risk here, because that's part of the experiment. I have no judgments around people reacting negatively. I was actually surprised at the number of positive responses or intrigued responses. Certainly there were people that were very freaked out and that were nervous. I got a lot of Instagram messages - very thoughtful Instagram messages - from people voicing their concerns. I love that. I love that I'm able to hear from all these perspectives, and that allows me to take those perspectives back into the work and also point out things that are necessary for us to think through. I was genuinely surprised by how many individuals were applauding AI clone Taryn, at least on my YouTube channel. You have to keep in mind there are a lot of people that follow me and have followed me for 10 years. I haven't been posting much. Maybe they're just excited that some version of Taryn is reentering the world and re-engaging on YouTube and they don't care that it's a clone version as much. Who knows the reason? But the response is pretty positive.
Mason: Do you think you'll clone AI Taryn to learn more about Taryn? So AI Taryn right now may love the Spice Girls, but what if you could create an AI Taryn that would just love punk rock? Might you realise that, oh, in actual fact I don't hate punk rock.
Southern: Yeah. I would love to have this virtual version of me out there, learning things and doing things. I was just looking at the AI version of Sam Harris, Sam Harrish. He's spouting all of this wisdom and then they injected some humour into him. I already think Sam has a great sense of humour, but this AI version of him was particularly funny. I love the idea that I could pick up some new skills. The AI version of me can pick up some new skills and hobbies, and and teach me some things. I'm into it.
Mason: Wow. An AI Sam Harris that is 20% Ricky Gervais. Well, the Digital Sherpa thing is appealing because that reopens this back up to the real joy and promise of the internet, which was, we would be able to explore and experiment with our identity. It feels like now we have to create these fixed identities or these brands: the Taryn Southern brand; the Luke Robert Mason brand that can only Tweet about science and technology. If I Tweet about anything else, no one cares. Creating these things to be Digital Sherpas on your behalf, to explore the other aspects of your identity that you don't necessarily want to discuss with your 400,000 strong audience - that seems to be the exciting and positive thing to me.
Southern: I like it. I think we need to clone you Luke, and then you can start running these experiments too.
Mason: I'm fascinated by the work of CodeMiko. Have you seen? CodeMiko is one of the first V YouTubers - virtual YouTubers.
Southern: CodeMiko?
Mason: CodeMiko. So it's created by an amalgam of technologies, but it's a virtual avatar that's fully puppeteered. So it's doing facial tracking. The person is wearing a suit that allows them to track their body. They basically sit on Twitch and have these interactions with their audience. Because this thing is an avatar version of themselves, because they're puppeteering it, they can express however they want. So the technician's thoughts - the human individuals puppeteering this thing - aren't necessarily the character's thoughts. I've always thought about it Taryn, and I would love to get to the point at which we can do this interview again, but it won't be you and I. It will be Luke Robot Mason and AI Taryn having a podcast interview. We could just let them go for 60 minutes and see what they come up with and see what weird and wonderful places they go into. Then we can kind of follow up on that interview. There's something beautiful about this stuff. As cynical as I seem, there's something beautiful about this stuff when it teaches us something about human creativity and human possibilities.
The idea of a CodeMiko where you can create new skin for yourself. I could be a green alien and have a completely new character and personality. I could explore some aspects of the thing that I wanted to explore as Luke Robert Mason, but the fact that I do science and technology stuff means that I don't feel that I can authentically do that online - that seems entirely liberating to me. That brings us back to the original dream of the internet to allow anybody to be anything they want to be online.
Southern: So good. I'm all in Luke.
Mason: Well we will have my giant green alien Luke Robot Mason interview AI Taryn in the future. The thing I love about you Taryn is you're always looking a couple of steps ahead. I know that you have this wide interest in things like biohacking and brain computer interfaces, and I wonder how some of that work might start enhancing the AI work that you're doing. Do you ever think we could get to the point where we could just jack in to computers to express our creativity?
Southern: I do. I do think that we will. I'm already amazed at what's being done in the brain computer interface space. What putting an electrode inside the skull already means for human connection, creativity, emotion. Of course, it brings with it a whole slew of controversial questions, but the reality is that we can manipulate human behaviour. We can shift human behaviour. We just have to decide what we want this human machine connection to look like.
Yeah, I'm super passionate about everything that's happening in the world of biohacking and longevity. I think personalised medicine and diagnostics are changing the game for what it means to live our best lives. I think us being able to hook these diagnostics into our storytelling and into our creativity is where things get really, really interesting. There's a job title that I really, really want called Experience Architect, where I'm essentially building experiences, via these story and sensory input sounds, tastes, smells, sights. I'm building an entire world that people can enter into and get lost in and embody storytelling. I think that's the future of creativity and storytelling. I can't wait for us to just get there already so that I can do it and have Experience Architect written on a digital card or on my LinkedIn bio.
Mason: Well, that's where we're going really. It does feel, talking to you Taryn, that the AI clone is just the gateway drug for you to create your own reality.
Southern: That's right. That's right.
Mason: Well at least you're willing to admit it. When we die, we can all decide to come live in Taryn Southern or Taryn Northern or Taryn Eastern or Taryn Western.
Southern: By the way, I don't claim my world is the right place for anyone to go and hang out in. We'll leave that to everyone to decide
Mason: We'll all make our own decisions. I mean, I guess inevitably this ends up in asking philosophical questions about what we believe the brain and human creativity is. There does feel like there's an underlying assumption in your work that you feel that the human brain is very similar to a computer or an information processing system. How true do you really think that is Taryn? Are we just hyper evolved computers, or is there something more intangible that's going on there? There's some other form of thing that allows human beings to express creativity.
Southern: I don't think we know. I don't think we have any idea. I will say from all the work that I did on 'I Am Human' - which was this documentary about the feature of the brain - I learned a tremendous amount and was really struck by how algorithmic we really can be in the early neuroscience studies, looking at how the brain assesses information and makes decisions. We are the product of millions of experiences and memories and how our brain categorises those memories and experiences to create learnings. That's an unconscious process that's happening in the background. It's like the processing of a computer. We don't know why certain pieces of information are given more relevance than others, but we can kind of quantify it and there are patterns there.
I'd say my inclination is more to believe that we are just incredibly advanced algorithms. We just don't yet have the tools to fully dissect what these algorithms look like. I know that some people find that terrifying. I find that empowering, because the idea that we really are just these complicated algorithms that can be reduced down and changed for the better. That we can take elements of our personality that aren't serving us - such as anytime someone says a negative comment about me - I'm going to remember that. But then for some reason I just forget about all the positive things. That's a kind of crappy flaw in my OS system. Can we shift that to create a more positive Taryn and a more balanced Taryn, and maybe a Taryn that's actually more reflective of reality? I like this idea, actually, that our brains are these advanced algorithms. It's soothing to me.
Mason: But aren't all Taryn's flaws what makes you unique and so interesting as a creator? It feels like an odd and difficult question to ask, but if you were able not to go through the experience of cancer that you went through, would you choose to? I guess from a medical perspective you would, but what you've learned going through that very negative experience, surely that's now changed your thinking about life, creativity, the AI, all that you do? There is value in these negative experiences, even if they are negative.
Southern: 100%. I can't argue that I think we should just wipe out negative experiences. I don't know how you figure out what's an experience that's worthy of keeping and holding onto, and what's something that's actually just holding us back. But even regarding the year that I had in cancer treatment, there were so many beautiful lessons that I learned in that year. Coming out of that, I kind of felt invincible. Nothing's stressing me out. I'm so grateful to just be here and be alive every day. As time goes on, you start to lose some of these insights. They start to become like distant little fuzzy dreams, and they're so important to be able to hold onto. So I would actually argue that I don't want to take away the negative experience. I want to fortify the learning and the lesson from that. What's a way that I could do that? Through future technologies with the brain. How are we going to make these decisions? How are we going to figure out what's worthy of saving and getting rid of? It's a really complicated, much bigger conversation, obviously.
Mason: Yeah. It's just wild hearing you say that. I was binging on Taryn Southern content this week in preparation for this. I was watching some of your older videos. Then there's the more recent video which is about surviving your cancer diagnosis. You see this individual is so connected with human beings, who is hugging them and cuddling them, or crying and expressing full emotion. It's such a different Taryn from the Taryn that I'd binged on your YouTube. It was a more authentic and more real human being. I'm not saying that you should drive towards more of that style of content, but it was revelatory in a way. It was a beautiful thing. It was a beautiful thing to see. It was in such vast contrast to your AI clone video, which was the next thing that was for me to watch.
So I guess my, my final question is - I guess it's a philosophical question - and it's one about creativity. Where do you think your creative expression comes from? When we talk about AI, we always fixate on consciousness and we fixate on intelligence. But in actual fact, it feels like creativity is not reliant on consciousness or intelligence in any way, shape or form. Sometimes it's reliant on dreaming or just being unconscious or going for a walk and coming up with an idea. All of these things that we discount in everyday life is where the seat of creativity is. So Taryn Southern, where does your creativity live?
Southern: Oh goodness. I feel like my inability to answer this question is probably the answer to the question, which is, I follow my curiosity and I just always feel like it's this desire to understand something that I don't understand that fuels my creative process. It's probably why I've always been driven to do the thing that no one else has done. I don't mean to sound egoic. Even when I went and started making YouTube videos back in 2006, it was not kosher at that time. There was sort of like, "What are you doing? Cat videos belong on YouTube, not people." I don't know. I just have this insatiable interest in doing either the wrong thing or doing the new thing or doing the thing that will just satisfy some deep urge to know. From there, when you're present with those learnings and those insights, they naturally evolve into some kind of creative process. For me, I get most excited and I light up when someone poses a question or I hear something new that I have never heard of before. I instantly feel like I've got to dive into that. I don't know what that is, but I've got to figure it out.
Mason: Well Taryn Southern, you're a wildly creative and a deeply interesting individual. I'm excited to follow the evolution - or devolution, maybe - of AI Taryn. On that note, I just want to thank you for being on the FUTURES Podcast.
Southern: Luke, it was so good to see you, it's been way too long. And I've got to hear all the amazing things that you're up to next.
Mason: For another time.
Southern: So thanks for having me.
Mason: Thank you to Taryn for showing us what's possible when code and creativity collide. You can find out more by visiting her website, Taryn Southern dot com.
If you like what you've heard, then you can subscribe for our latest episode. Or follow us on Twitter, Facebook, or Instagram: @FUTURESPodcast.
More episodes, transcripts and show notes can be found at FUTURES Podcast dot net.
Thank you for listening to the FUTURES Podcast.
Credits
Produced by FUTURES Podcast
Recorded, Mixed & Edited by Luke Robert Mason
Audio Editing by Elliott Roche
Transcript by Beth Colquhoun
Social Media
Twitter: @FUTURESPodcast | #FUTURESPodcast
Instagram: @futurespodcast
Facebook: @FUTURESPodcast
