 So we're here at the Lunara Connect Hong Kong 2018, and who are you? I'm Laura Decker. And you just did a presentation about your, is it art, did you do? Yeah, yes, I actually have a technical background so it takes me a while to realise that yes, actually I now work as an artist. So it's like art that uses AI or what is, how does it work? Yeah, most of what I do is using AI at the moment. I've been producing machines, kind of sentient machines. So for example, this was one of the slides that you were showing at the, so there's like projector poetry on the floor and what's going on here? Okay, so this machine which has kind of constructed itself from discarded components and junk is grown out of the junk and it's beginning to explore the world that it's found itself in and so it's looking, using its camera, looking around at what it can see and it's tasting it. So I've used a neural network, I've trained a neural network so that the machine can transform its video stream into some sense of taste at what it's seeing. And then you were also showing some little robots that you had, you have some AI robots that are like having some kind of role? Yeah, so I was noticing that the media were presenting stories about AI that were quite hyping and instilling quite fear into people so I was collecting headlines, I was quite interested in the way these headlines grouped themselves about the kind of the anxieties that people had about AI. Do you have any anxiety about AI? Because I think it's like overhyped, like there's not going to be, it's not going to take over the world. I wouldn't be surprised if it does become quite powerful independently. You would or you would not? I wouldn't, I think it might well do and I think, so I think there could be quite a bit of confrontation in working out how to sort of control if the AI becomes quite independent and powerful. But you said you're not too worried that the AI is going to take over your job but over here it looks like it's doing all the poetry and the like the graphics and stuff so you're not worried that the artist is going to be disrupted? Yeah, I think that actually that there's 4% likelihood of my job being taken over. No, I think that's probably a way under estimate. So 4% per year? So that means within 20 years is for sure or what? Oh I see, that's a very good point, I don't know. It wasn't my, it wasn't my, it was just a BBC website that was doing this. So you know Kantar from the age super computing Leonardo, right? You were a student together? I wasn't a student, yeah Kantar was doing her, I actually only was an undergraduate too and then she was doing her PhD. I was working as a researcher at University College London. And she's doing some super computing stuff. We were both working on genetic algorithms at the time. What is genetic algorithm? It's using kind of the evolution as a model as a way of developing better solutions to a problem. So life, the problem of living and existing and feeding and so on. We have our genetic code. The idea of genetic algorithms is that you represent the problem that you're trying to solve as like a genetic code. And you have a pool of candidate solutions to your problem that you set free to mate and reproduce and they eventually develop better solutions to the problems that they've been set. Was it something that Alan Turing studied after he thought of the computer and stuff and then he wanted to do stuff that's related to that or? Evolutionary computations. Yeah, is it something, no? Maybe. I don't remember hearing his name associated with that actually. So the probability of life going one direction to the other, is that what it is? Again, no, that's slightly different, I guess. No, it's using genetics as the kind of the mechanism of it, but also you can think of it as a metaphor, whatever problem you're trying to solve, be it like chip design or living and finding a mate. And then here you have a graph showing the whole process of what you do? Yeah, so the idea is quite heavily influenced by the work Daniel Dennett has done on the human mind and consciousness, the idea that the brain doesn't have one centralized control, there's actually lots of processes all competing with each other, sensing things, producing outputs and then kind of almost competing with each other and what you choose to do, what happens is a kind of net result of all of those processes fighting. And what have you heard about Delano and all these engineers that are doing the open source crazy stuff, making the whole world work on free software? It's wonderful for me, that's what I use, whenever new stuff comes out, it's really exciting for me because it's new tools that I can use in different ways, so I'm massively grateful for people who do that. Big announcement from the Leonardo 96 boards, they're launching the AI boards, they have a whole bunch of new development boards, they have these neural processors that can do faster AI and so you might enjoy playing with those. Totally, yeah. And putting some algorithms on there, I guess. Absolutely, at the moment I do all that stuff in software but hardware would make it possible to do faster and do more. You might get surprised but what comes out of your next art piece, right? Yeah, totally, yeah. Right, and where is it going to be? Where? Yeah, you're based in the UK? I'm based in London, actually I'm working on something for a Discovery Museum at the moment, I make it, it's for a kids' Discovery Museum. In the UK? Yeah, in London. No, it's in Halifax, just north and I'm making an interactive sofa so when the kids stroke the sofa things happen. I want that at home. I think my sofa needs to be smart, I'm waiting for the smart chairs, smart floors, smart everything. Yeah, yeah, well I've done this before, I did a version of this for a different exhibition that I didn't have time to talk about but as this psychotropic lounge, so I made this sofa that you stroke and it changes the wallpaper while you're doing that. And kind of Sony has been in difficulties kind of like with their money and stuff like that but recently they showed a little dog, like an iBow dog and it supposedly has an AI and it's like I'm not sure what it does, you know, but I think that could be, that could be a future for a lot of big tech companies to do things that doesn't, they don't have to be like completely functional or something, but it gives you another nicer feeling to be interacting with some kind of AI. Yeah, they're actually been those out for quite a few years, in fact I got some for my niece and nephew and they have what they call Naughty Mode which is presumably the machine just breaking down and not functioning but is turned into a virtue by calling it Naughty Mode. So you could be maybe consulting with giant companies to make them launch some super nice AI sofas maybe in mass production? Yeah, absolutely, yeah, yeah, yeah, yeah, some top-end furniture companies, that would be great, yeah. All right, Aki, I hope they're watching. Okay, cool.