 Hello everyone and welcome to this beta's own session on AI and robotics meets human creativity. I'm Bert Xi, moderator of this session where Su Wencheng will speak for about 15 minutes followed by time for questions and answers for myself and the audience. As a professor of electronic and computer engineering at the Hong Kong University of Science and Technology working on human robot interaction for healthcare, I'm very excited to moderate this session with Su Wen on the interaction between humans and robots for art. Su Wen is an internationally acclaimed Chinese Canadian artist and researcher based in New York. She's widely recognized for her pioneering work in the area of human machine interaction. Her work explores the made by hand and made by machine approach to understanding the dynamics between humans and systems. Su Wen is currently artist in residence at Bell Labs and at the New Museum of Contemporary Art in New York. She's previously held positions at the MIT Media Lab, Google, iBeam, Japan Media Arts, and P9 Autodesk. She has spoken frequently on artificial intelligence, creativity, robotics, art and science innovation and interdisciplinary collaboration. Her TED Talk on post-human collaboration has been translated into over 20 languages. Her work has been exhibited in museums and galleries around the world including Tokyo, Singapore, Cambridge, Massachusetts, New York, Geneva, Montreal, Mexico City, and Barcelona. She's served global recognition for her work, most recently winning the Fallen Walls, Science in the Arts category in 2020. Please join me in welcoming Su Wen Chang for the beta zone. Thank you Bert for this kind introduction. It's a pleasure to be with you here today. Two sets of eyes take in the same scene. The same light passes through two different apertures that see entirely different things. One set of eyes is muscle and nerve, evolved over millions of years. Unique in the universe, perhaps, but better understood with each passing day. One set of eyes is metal and plastic, engineered for the task at hand. Uniformly constructed yet growing in a complexity we may not be prepared for. One is mechanical, the other biological. And they collide and this collision might seem extraordinary and it is, but it's also the basic condition of our lives today. The truth is we're seeing double. It's the double view of machine vision, the view through security cameras, social media and satellites. We've constructed a world where all sight is dual sight and all seeing is a bridge between the contradictions of nature and machine of art and engineering of self and the collective. The poet Adrian Rich said, we're living in a time of unprecedented complexity. Our senses are currently whip driven by a feverish new pace of technological change. The activities that mark us as human, though, don't begin, exist in or end by such a calculus. She said that over 20 years ago and that's only been more true over time. It's true. The rate of technological advancement is accelerating. The machine's role in society is evolving. The questions raised by AI systems, robotics and technology can point to a bleak future. They make us ask questions about whether machines will replace robots and whether machines will replace workers and whether machines will replace artists too. It's hard to look at a headline today without feeling one's anxiety stir. Will we be replaced? Will we replace ourselves? Or will the machines somehow save us through technological advancement? Will they liberate us from want, keep us safe and healthy in a way nature never could? Will they offer us a way to transcend our evolution? This duality, the promise of its power and anxiety of its consequences, shapes my worldview. I would go so far as to say it shapes us all really, with every time we're thrilled to connect with friends around the world and every time we're convinced our phones are listening to our private conversations. With every time we throw a prompt into mid-journey and then wonder about the training data. With every time we get excited about the release of a new phone while seeing images of technological waste in landfills. My work at its simplest level is about exploring the contradictions that stem from duality. Exploring how these contradictions point at a third path, what I like to think of as a new hybridity. It's a space I feel at home at. I think it might be because I'm a product of two cultures, of two languages, as a child of a musician and a computer programmer, with the fascination with science and art. I think that's why I've spent the past decade compiling the oldest forms of art, with new forms so advanced we actually have to invent the technology as we go. And it's no secret that you and I and everyone we know is a participant in these contradictions as well. Often we live this duality unconsciously and in our sleepwalking we risk forfeiting our agency. I explore these themes in my work so I can feel grounded within these contradictory poles by actively building bridges between them. We're already living hybrid lives so let's make choices about that hybridity. So I look to art to serve as a bridge between these contradictions. Not to resolve them into some new cohesive whole but to entangle them and illustrate the ways in which they're already entangled. Paintbrush and pixel, art and engineering, human creativity and artificial intelligence. It's about engaging in an active process past techno-optimism, past reconciliation, exploring contradictions in the natural and the mechanical, exploring the medium of art itself. I believe that art is a process of thinking and these processes are rituals and rituals, old and new, are vital forms of world building. A way that fear and hope can be held in the mind at the same time. Machines treated with compassion yet imbued with perceived agency but understood as the product of human bias and not peers in thought. And to bridge these contradictions we need to build more syncretic disciplines and comfort with practices that unify multiple traditions. Working with AI and computation is a way to expand what art is and what art can be. Art as cognitive science exploring alternative forms of consciousness. Art as engineering, constructing robotic systems for empathy and connection. And art as philosophy imagining new social and moral constructions for this human-machine relationship. And at the intersection of these fields my team and I create. We create simulations, performances and artifacts across scales. We work in the large scale like this relational robotic performance integrating satellite and biofeedback. And in the small scale, like these, like the small scale, like these prototypes for robotic units inspired by the microscopic threads that link fungi into a biological network. We look for intersections, that gray area between human and machine and we stumbled upon it in the first project in a series called drawing operations. The practice began with Doug one. Doug one was about scene which is the first step towards knowledge. Because to know the world we must first observe the world. It worked through computer vision and a robotic arm. We started by taking one of the most visible frequencies in the color spectrum red and deploying it for our first collaboration. We thought if we could utilize the machine's reductive functionality, its ability to be selectively colorblind, we could communicate drawn gestures to the robotic unit in real time. By converting the movement of the drawn line into digital signals, we were able to suspend those lines. Like recording the choreography as an improvised dance as it was unfolding. In order to create a duet on canvas, we relayed the positional data back to the robotic unit, back to the stream based simulation, back to the space of embodiment. I aimed to break the dichotomy of human versus machine, to build a technology that doesn't encourage more of the same or reduce through replacement, that doesn't atrophy physical processes in service of computational ones, a technology that creates something new as it's being used. After exploring the duality of sight of computer vision versus the human eye, we were inspired to explore the next generation of drawing operations. We moved from sight to the record and recollection of sight. We were inspired by memory. We looked for the possibility of approaching machines beyond pushing the physical limitations to parameters to its limits. To do that, we needed to draw from a larger network and construct a framework for machine interpretation. So like human memory, we looked at two decades of my drawings as a kind of data set. It was like embedding a record of my own memories into the artwork. We trained a recurrent neural network on two decades of drawings. It produced an AI system rooted in the analog and given new life in the digital. Memory and acting on memory created a collaborative partner in Doug 2. This revealed the possibility of using machines not only as tools, but as non-human collaborators. Now I'm not in control. I'm working with it to a common and unexpected outcome. It leads and I follow and I lead and it follows. It's a tree of decisions based on uncertainty, flow and improvisation. We're being told that the purpose of art and AI is to automate the process of making and I disagree. Instead, I'm adapting my own artistic gestural memory through working with algorithmic prediction rather than being replaced by them. Automating my work isn't the target of the AI system. It could be said that the approach to art is the approach to co-creation. To manifest this third thing, it's co-equal and gesturally empathetic. It points the way towards a relational interaction that offers a new form of knowledge production in action. Art practice as research, art as interdependence and not co-dependence. With my third drawing operations piece, I thought of the gaze of the machine and began to see vision as multi-dimensional, as views from somewhere. We collected video from publicly available camera feeds on the internet of people walking on sidewalks, cars and taxis on the road, all kinds of urban movement. We wrote a vision algorithm based on a technique called optical flow to analyze the collective density of velocity and dwell states of urban movement. Instead of a collaboration of one to one, we made a collaboration of many to many. By combining the vision of human and machine with the city, we reimagined what a landscape painting could be. Sometimes I wonder if perhaps the future of creativity isn't in what it makes, but how it explores new ways of making. Not either or, but and. Not binary but plurality towards a continuum of relations. Because doing this work has taught me how abstractions like artificial intelligence don't come from nowhere. They're constructed through our existing histories, philosophies, and cultural perspectives. Our metaphors for the other. And they don't exist in a binary but in a continuum of material, social, and cultural relation. It's easy to forget that there's no such thing as artificial intelligence because there's no such thing as a single natural intelligence. If we continue to challenge our idea of what intelligence actually is and what these configurations with technology can be, what would we be able to see? In my view, this is where we are today. We're seeing double. It's not human versus machine, but human and machine. This is our future. And in many ways, it's our present. To me, this is our emerging relation between human and machine. This is the third way, the hybrid, where human intelligences and the unique agency of machines collaborate. This is my bridge between the contradictions. And I'm excited for the beautiful unknown wonders that will emerge from it. Thank you. While we're waiting to set up, thank you very much for that speech. It was really inspiring. I'd like to maybe take my rotteries for privilege and ask Suen a few questions before we open it up for the floor. Suen, in your work with Doug Won, you've mentioned that the limitations of the robot actually were part of the adaptive process. I'm thinking as a technologist that the limitations of the robot and its ability to rough-cake human behavior may shrink in the future. Do you think that will affect the way that your interaction with the robot will happen and what you can learn from it? Absolutely. I think with each generation of robotic unit, I'm on my fifth generation now. We've explored the limitations of the system and kind of tried to push against what that might be. For generation one, we worked with simple computer vision. With version two, we used a recurrent neural network and delineated the drawing data from two decades of my analog drawings. The limitations between analog and digital there. With generation three, we took the crowd data from New York City and implemented that into a system that extracted the positional gestures. I think we're always really quite inspired by the limitations in order to create something expressive. Also, in terms of understanding the data, understanding the technology a little bit more. I think with each generation, I've learned a different thing. As robotics and AI become more advanced, I think there's always going to be points of friction between what happens in maybe the analog and in physical space and what happens in the simulation. I think that's what we're always trying to push against. I think through developing this kind of expressive configuration with machines, we're able to unlock different novel use cases for the integration of robotics in society. Robotics with a sort of human behavior and gesture. I think there's a lot of different kind of entry points for thinking about different kinds of sensor development in this way. I'd like to pick up on what you said about the interaction between society and humans. Part of your work you mentioned picks up on the duality between our hope for the future and the power that AI and robotics is going to bring us, but also our fear at the same time about the potential consequences. This is a theme that we've seen in many other of the sessions, but kind of viewing it from a different perspective, let's say technological or economic or policy-wise. I'm wondering, from your perspective as an artist, do you think there are things that the art perspective of this interaction between humans and AI can bring to the table that those other perspectives might not have? Yeah, absolutely. I think obviously what I said in the previous question about developing novel use cases and features for different types of technological development, that's maybe kind of the more engineering focused mindset, but I think at the end of the day what interests me about this approach to art and technology and art and engineering is it's about a meaning making and about trying to create a novel and maybe more existential problem from the tools at hand. I think there's a lot of value in artistic practice from a level of creative expression, but also trying to make sense of a lot of these new developments in image generation, in sensor development, in a way that can really impact the individual. I think artists throughout history have always attempted to create meaning from their tools and surroundings and I think this is no different. I think the difference with this is I think there's an opportunity to really explode and really illuminate the black box of technology and make it meaningful for the sort of maybe the person who's not in industry and not deeply involved in the development to understand their role in that configuration, so I think there's a lot of interesting intersections there too. Yeah, I really resonate with your point about creation of meaning because I think one of the fears that a lot of people have, you know, face the AI is kind of a loss of meaning, right, because you're taking over your job. There's a lot of misinformation as well. And things like that and so I think your emphasis on using this and our interaction with it is a new way of creating meaning, it's a really powerful metaphor and way of thinking about this. Yeah, so I don't want to totally hog all your time. Of course I could talk to you forever about this, but maybe we can open up to the audience if there are any questions. Just raise your hand. Thank you for the very impressive presentation. Also as we can notice from the video and there are many paintings that are completed by you and AI, you call it as core creation, yeah? And how can we define the values for such creation or other works completed by the artificial intelligence and also human as the core workers? And the second question is, as we can see, there are more and more AI artists in different areas including music, paintings, but it can be an opportunity and also challenging for human. But can you use several words to summarize the relationship and the future between human and AI or missions? Yeah, thank you. Yeah, I think there's quite a few questions there. It's something that keeps me up at night and I think about a lot, because on one hand it's really exciting to see so many people engaged in visual culture in a way that hasn't been available. I'm thinking obviously image generation like Dali, mid-journey, stable diffusion, all tools which I've tinkered with a little bit. Obviously that's really exciting, but on the other hand it's that promise and peril of these tools and that does exploit, there is some concerns about attribution of artists in those data sets and what. And that's obviously an area to be taken very seriously. At the very least it displaces a lot of industry and jobs. So I think for the industry of image generation we're witnessing a lot of new policies unraveled that I hope will not stifle the incredibly powerful possibility of people being involved in visual culture without exploiting visual practitioners. And I think there is, there must be an ethical way to approach that if we can solve for so many things we can solve for that hopefully. For my work which thinks a lot about gestural memory, artistic memory, and feedback loops and time, a lot of what I do exists in physical space. So a lot of what I'm doing is adapting to the robotic unit through various sensors that record my presence in the room for the simulation and then the deployment into the physical apparatus. But I think for me the feedback loop is really mark by mark to the point where that delineation is not as important as the thing that is made together. That's the metaphor and the process that I use. I think it's really exciting. You see these types of feedback loops deployed in robotic units being used alongside surgeons, right? It only enhances the practice of that delicate industry. So I think my work is maybe, maybe I should never be a surgeon, but I think it's more along the line of layers of human-made marks and machine-made marks that really drive what I do. So when I think about collaboration I think about whether both can be implicated in a process that I, you know, both learn from. I think you've seen through almost 10 years of this robotic development that it's continually in motion and continually building upon the next. And I think I would say that my own practice has as well. Yeah. Can you wait for the microphone? Thank you for the presentation. With your experience on AI and creativity, what is your thoughts on AI and filmmaking? I think some of the new image style transfer, I'm thinking runway generation three, I think it's really exciting actually. I think there's so many ways that AI filmmaking makes you think about style and memes in a new way. I'm thinking about the Wes Anderson Balenciaga edits. Again, trying to find new layers of meaning with traditional filmmaking and that kind of narrative is, yeah, I think it's really surreal but very, very exciting. Again, I think it's about getting more people into the language of filmmaking, into the language of narrative and time and video. So at the same time, obviously there are some concerns, obviously there's the writer strike going on in the United States about not eliminating the human in the process and also being respectful of what industries are displaced. So, yeah, it's always sort of to head a dragon, I guess, but for my purposes, being always fascinated with film, I found it to be a really interesting entry point. And to refer to older artistic styles and animation that really blurs the line between live action and the image. Yeah, I think there's a lot to be said about that. Yeah, because I recently saw Tom Cruise with AI, they had Tom Cruise in the next Mission Impossible. It's going to change a lot of filmmaking. Yeah, definitely. But I also think that using Tom Cruise as an example so much about what he does is about the narrative of the stunts that he does, the narrative of the human actor pulling these incredible feats of athleticism and daring that always draws to the box office for his movies. So that can't be replaced by AI systems unless we are interested in an AI system jumping out of a plane at any point. But, you know, I definitely think a lot of the ways in which the technology is innovative becomes quite sensationalized. I think there's still so much room for human presence, even in the use case of a Tom Cruise movie that won't be replicated anytime soon. But we'll see. Oh, right in the back here. Good morning. I wonder if you could just talk a little bit more about your co-creation concept. We tend to think of tools as an extension of human capability, but you're implying a creative feedback loop where you're seeing creativity originating the machine, which is feeding back to you. Could you expand on that a little bit? Yeah, definitely. I think I can understand that co-creation is a term that gets thrown around a lot these days. What I find co-creative about the work with the robotics in particular is that with every interaction, with every time I engage with the system, I have to actually physically learn how to work with the output of the recurrent neural network, physicalized. I think there's that aspect of it that that makes an end the fact that my gestures are in response to gestures being made on canvas. There's so much of that that it's how I would define collaboration, which is mark making their time, and both actors being creative catalysts for one another. How that works on the machine side and we're actually developing a new generation based on this is that we're inputting new data into the next generations of drawing with Doug, drawing with drawing operations at the unit all the time. In that way, there's a relationship between how I adapt and how the machine system responds to my work that for me hints at co-creation. I will say it's not the same type of collaboration as a human being. I think that's a false metaphor a lot of the time, but I do think there's something very tenuous and time based about the practice that keeps on invigorating my own artistic and creative and technical development. We're pretty much almost out of time. Maybe I'll just use my moderator privilege to ask the final question to wrap it up. I'd like to ask you about the future. You talked about a number of systems that are integrating more and more data and different types of inputs and so what is the next step in this journey of your exploration of the collaboration between human and machine? Thanks for setting that up. So we're obviously there's so many interesting ways to approach this intersection. We've started, this is how I like to see it, we started from the very small and immediate. We've started from the drawing data set where most drawings are very much in the realm of art, very small in scale. We've looked at computer vision. We've looked at movements of cities and from that we went to the very small, you'll see here, I'm painting with biofeedback EEG headset. So we went from the city to the scale of a brain wave and I started thinking that in all these processes it's just a series of flows. What flows through the robotic unit is electricity. What flows through my own system is electricity in a different scale. So I've been really interested in thinking about planetary flows and what that means for thinking about the earth as a cybernetic system, which it has some precedent in certain areas of philosophy, thinking about Gaian systems and such. So yes, expanding the sensor apparatus to the scale of the planetary to try to understand a little bit more about what that means for the next generation of the robotic units, which is called flora rearing, agricultural network, Fran, a series of robotic units that steward nature powered by a sustainable resources like solar and microbial energy. So yeah, breaking apart even what is a machine and what is an input is always a source of endless curiosity for me. So I think that's next. That's really inspiring. I'm sorry we don't have more time to follow up on that but I think this idea of like integrating art and humanity and actually the entire planet together is a great note to end on here at the World Economic Forum. So thank you all very much. Thank you. Thank you so much. Thank you very much.