Voice-Assistant Extended

Conducted in English

Adrian Demleitner 0:02
Good. Thank you very much for taking your time for this briefly.

3j14j0 0:06
It’s okay. No, it’s nice to, it’s nice to catch up with you also see what

Adrian Demleitner 0:13
I mean, I, I attended your workshop last year within my master studies, so I’m doing the mastering design research. And I was very much interested in our relationship to technology from a non human perspective. You know, a bit but you’re trying to do with unfamiliar, convenient. And I was, I started to focus on voice assistant technology at that time, and your workshop was a really good start to, to explore that rhythm. So I, I mean, my, my personal inspiration is, you know, stuff like vibrant matter, or the modern human discourse.

3j14j0 1:44
Vibrant matter. Actually, I’ve never heard of it. So what is the Vibrant matter?

Adrian Demleitner 1:51
so vibrant matter is a bit like the Eurocentric philosophical approach to animism. Okay, you know, it’s a bit like, what’s the political dimension of ascribing agency to the non human realm? It’s a bit it’s a, it’s quite philosophical. not always easy to read. But it’s a it’s good. I can recommend it. I’m just taking them off myself. Yeah, that’s good. Do that. I’m always I’m always available if you need to have a link or something.

3j14j0 2:31
Yeah, sure. Sure. So yeah, I got it. So So you had to you had to kind of a bit pivot into into basically, into more like, a social thing, because it’s like, it’s, uh, you know, as like, I see how more than human perspectives are socially relevant as well. So but yeah, but I guess?

Adrian Demleitner 2:55
You know, so it makes a and I think, in communicating the nonhuman relevance of these approaches, speculative is very important, because it’s kind of the language to, to bring the imagination that it’s needed to understand this. Sure. But yeah, it’s, it’s, I will keep that part, but I will continue. After my studies. And for now, I focus on something called the post phenomenal post phenomenological approach. Okay. And that is, while phenomenological approaches are focusing on how experience through the senses is creating access or reality, like how you see stuff, the past phenomenological approach is focusing on how technology mediates that access. So for example, how are glasses changing the way you have access or perceive reality? So that’s a bit where I’m at right now I’m focusing on how to voice assistance, create or mediate access to the world today, and of course, that has political dimensions with the whole data extractivism thing and the business model of PMF of institutions like Google or Amazon, but I will also go a bit into the social relevance of voice assistant technology for people with with with disabilities, who are dependent on human computer interfaces that go beyond the usage of eyes or hands. So, yes, that’s a bit where I am right now.

3j14j0 5:13
I think there’s like, so, I think superflux they did like another because this one that we showed you that the, the our friends, sorry, yeah, our friends electric, which is that main film with Mozilla. Exactly. He did another one, which was specifically maybe for any chest, which is the, you know, the national health security. And so they did like a number one, and it’s also about like device assistance. I’m a bit less kind of, it’s a bit random, I think maybe they just agreed, and then they did it. Because I it’s like, it’s a little bit confusing. Okay, that’s a kind of maybe some of the things like fall into topic, you know, of the interesting. I mean, I get, again, where you kind of go in and like I was reading also, I was a bit curious, because I’m also a little bit interested in these new new use cases also. And then like I was reading about how, well you know, they introduced the Alexa Pretty please, like scale for kids, because they said that, you know, kids were really like impolitely addressing Alexa and stuff. So parents were kind of a bit shocked to know that the election was like slavery almost right. And it’s interesting, because they Somebody done again, somebody did a study, and then said that actually had no sort of influence. Now not clear for some kids, because kids apparently are able to distinguish, you know, quite well, what is a machine and what is like a parent. But they said, but they said basically that, you know that there was this interesting point, which again, I guess falls into your post phenomenology is that it is hard to estimate the impact of those kinds of things, because the kind of the social norms around those things kind of evolve as those objects evolve as well. Right. So that’s an interesting sort of perspective, you know, that it’s kind of the language is still developing, you know, and made me Yeah, so anyway, but but so you’re like, so people with disabilities, right?

Adrian Demleitner 7:32
I mean, what I will do in that, I will not propose anything in particular, but I For starters, I will have some interviews with blind people or people suffering from data or plugging pledgie, you know, when they’re numb from their, from the head on down to see a bit what, what, what their experience is or what their wishes are. Because I see, I see a big change in voice assistant technology. So we have the screen readers. And screen readers are actually more of a workaround to work with content that was created for the able bodied. So screen readers read, try to read off the screen. But the HTML, for example, was was made for people who are actually working with their eyes. Right. And I see that invoice the system technologies like like, like, Amazon Echo, that it’s a complete change in that, you know, its content made for the hearing first. So the workaround, and that brings, brings a lot of problems with it. And of course, Google and Amazon, they’re, they’re operating within their business models. So for example, they can brutally curate results in their favor, you know, and I don’t, so I researched it just a little bit, but I don’t think that Google and Amazon for example, have have other use cases in mind, you know, they are like, Okay, so we have this now, how can we place a lot of these devices in people’s homes, and how can we operate within our business model? So for example, Google has no not much information or resources on using this technology. With with disabled people. Amazon does quite a few things. They have a very well resources, but I just made a quick require inquiry with the Twister sales Association for blind people. They said they actually have a lot of problems setting up an Amazon Amazon Echo To start with, because there is a process of setting up and it doesn’t have an audio capture. You cannot even set it up.

3j14j0 10:01
Once you once you set it up, it’s like, Okay, what the answer is? Yeah. Interesting. Yeah, that’s quite funny. Yeah. Maybe related to Yeah. Interesting. Yeah. But usually, I mean, like, that’s the way actually the format or the framework is quite flexible. So in a sense that, like, every time is basically nobody told you at Amazon that they need to do it, you know, because as soon as you tell them, they will patch, they will patch it around and stuff, you know, that’s like, it’s a bit like, I think that it’s what’s interesting, really, with, with voice assistance from like, a very technical perspective, is this idea of a skill, you know, they really like it. There’s, there’s basically, it’s, it’s so modular in itself, that I’m like, I’m not boasting like that, suddenly, you know, big tech companies have gone on, and it’s like, it just applies to the general framework with all devices systems, is that, you know, open source or not, is that that the idea of a scale was just kind of makes it modular. So if there’s a problem, you basically solve a problem with the scale, almost, right? Like these, these things are set up and stuff that’s like a little bit more of a complicated, you know, sort of, well framework, because you still need to, you know, get it to work somehow. Yeah, I know, I met my friend, actually, my friend here in China, his, his dad is blind, and he turned blind over time. And then he actually he kind of became a bit obsessed, because not only he was blind, but also a little bit lonely. And he was he used to be like a video editor. So for him, like going blind was like, just completely, you know? Yeah, it was a disaster. But he became very passionate about technology. And so he has like, three or four different voices systems, you know, and like, a couple of them from different brands, and just kind of placed, you know, in these different locations, and then he just talks to them. And they, they activate his, like, free vacuum cleaners, you know, it’s kind of a thing, right? So he’s really happy about it. stuff. But But yeah, it’s, it’s interesting, anyway, yeah. Cool.

Adrian Demleitner 12:19
So I have, I had a very good read, again, of the unfamiliar, convenient page. And I am very interested in the project, the whole project, but especially right now, in your first case study. So could you tell me a little bit about the history or the process that led to your decision to work with this technology?

3j14j0 12:47
Yeah. Um, so I mean, actually, just to the site, the website has not been updated, because now there’s basically the two two case studies already. So So and they’re, they’re kind of CO working together. But then the idea behind it was that I think, um, well, my work generally focuses on the domestic space and kind of in relation to technology. And, and when we were talking with Claire, basically, we were just having this discussion that the current vision of a smart home in all of its, you know, utilitarian sort of approach, you know, an application airy approach, you know, has has sort of substantially failed to live up to the, to this standard of smart, you know, what, do you define a smart? And, and it’s not? Like, it’s not that it’s a question of time, like, the conclusion we came up with was that it’s not really a question of time, you know, not, but more a question of attitude, you know, so that’s where all this non human side, you know, come comes out as that, you, it doesn’t matter if technology’s really good or not really good, or if it performs well, and not just the idea that you kind of place these objects in this limited kind of slave capacity, you know, is sort of forcing the device is to do a behavior, which would be mimicking human, you know, right, and therefore, and therefore, that behavior, it’s like, as, as I say, you know, if now, you asked me, you know, to, I could use a vocabulary, and I could use all of my resources, you know, and then I would have to talk to you now, I’m in Swiss German, basically, right? So So the idea is that I will be in this constant, you know, Lost in Translation and kind of consumption of time and resources. So the idea behind this project was a bit to like, explore the native language of these domestic devices and see how they, you know, how they can inhabit the space differently and see it differently and whether that would open up You know, some sort of not exactly like novel applied to use cases, but more certain, like points new points of inquiry. So it’s not a lot about producing new outputs or sort of new configurations, but more about kind of opening up those devices to to new behaviors, and then observing how those take place. And, and then later on, if we see that there’s, like, you know, some interesting insights in there, we can like, think of some sort of more applied scenarios within the now it’s like, it’s half a research project and half a sort of almost artistic inquiry visibly, like we, we like make them you know, at the moment, they’re like, making poetry and stuff, you know, diverse and just like, making poetry based on where the vacuum cleaner goes. And, like this kind of

Adrian Demleitner 15:51
that was also be the question that I had, because it’s always a bit interesting to see where design is going, you know, is it more towards an artistic approach? Is it more, you know, factual research based? And I’m, I’m very fond of this kind of prototyping where you do something and then you reflect on it. And then you see what’s the learnings? And how can you move on from you? Yeah. So I was wondering a bit what your so you refer you rephrase the non human? And do you? What’s your inference there? Do you have like, some kind of theoretical base that you use to think on top of that, or? Yeah, so

3j14j0 16:43
I mean, the main, the kind of the global references are sort of the quite common ones. And so we’d like Take, for example, the object oriented ontology, you know, so the idea that an object is a sort of a part of the Anthropocene layer. And so that’s like, one of the things is like that. Yeah, exactly. And this Yes, yes. bogus is pretty. Yeah, that was one of the essential ones as well as that. Yeah, that just that day, did you know devices kind of have their own sort of ways of operate, operating, basically. And then yeah, I guess that that Yeah. That’s the main, that’s the main sort of theoretical framework. So like, not yet phenomenology, not yet. Both phenomenology, but and also ontology. And these two, and also, I mean, that there’s the, you know, there’s the Steelers side, which is talking about, you know, the theory of the machine evolution, basically, or evolution of machine as a species. So that’s, what was the book something something it’s like, an it’s like, a very long essay thing and split into four parts, I can find it the reference, but basically, yeah, so so he’s talking about this, this idea of, of machines, you know, that that are kind of developing also, as a as an independent species. It’s like a proposal of a theory. And then there’s quite cool. There’s this book by one Mian that I, I cite quite more and more, it’s, it’s called domestic appliances in post Mar China. And, and it’s like, it’s quite an interesting one, because he’s basically he is, he is like a researcher, you know, a theorist, but he would this is a bit of like, a side interest of miasma. He just like, sits at home, and like, looks at devices, you know, and sees and just kind of looks at how they behave. And, and he has this interesting, so he, it was written maybe 10 years ago. So he’s mainly talking about, like, more conventional things, like the lightbulb or the washing machine, or the computer. But as you can, it’s sort of rather poetically written so you can kind of, you know, take all these different sort of interesting aspects from there, which say, like, you know, a washing machine is kind of working in the language that we humans don’t understand. And it just quietly, like, spins, they’re, you know, occupying, it’s on purpose of some sort, but it’s still somehow, you know, it still performs whatever it needs to be performed. But you know, but we don’t necessarily understand what it’s saying, you know, we just, we just care about the end result and stuff. Right. So, so that’s like, that’s a bit the, I guess, the sort of theoretical foundation for it.

Adrian Demleitner 19:34
And I, I don’t know if you know, this one. I am a big fan of it. Every everything is so nice.

3j14j0 19:41
Yes, sir. Sure. I mean, we we even, we even drop like I think we throw it as the reference during the workshop already in the new showed itself. Yeah, yes. Mine is a friend of mine. He was actually based in Shanghai before. Not sure quite well. Yeah, I haven’t. Anyway, we’re, that’s pretty much like, that’s the kind of concept behind, you know, behind this whole approach, I don’t know what you know, in the end, it’s like hard to, you know, I say it’s, it’s interesting because it’s already hard not to anthropomorphize, you know, any of those devices, because even when you’re trying to come up with new behaviors, you’re still kind of, you know, taking inspirations from the, from how humans behave, or or, you know, or how sort of we observed some of these or something’s behavior. So it’s really hard to kind of really distance yourself. And and also, especially if you’re working on something that has to be visually represented or like that, you know, then it’s, you know, it’s, it’s hard to not think about also the emotions for for humans that, you know, this particular thing would generate or produce, or, like, how would the human Look at this, but at the same time, it’s quite like it? Yeah, it’s interesting, because the sort of less hierarchical things and yeah, I don’t know, I don’t know how to, like properly put it, but like, it’s, it’s interesting to look at, at at the home, which is sort of on equal terms with you, you know, then I think that there’s like much more of respect between different entities that kind of are, you know, inhabiting it and as I say, just because human made, it does not mean that, you know, easily, like if humans evolved from a unicellular organism, you know, then it doesn’t mean that these technologies primitive, you know, you know, but but when humans were primitive, and a bit stupid, nobody, you know, forced them to mimic, like, you know, I don’t know, whatever animal or something, you know, they could just be a human right, and evolve in their own way. So that’s one thing. And then the other thing is specifically towards assistance. I from this ontological perspective, what’s really interesting to me is that, um, well, it’s like, it’s the first object that even if forcefully, you know, is somehow sense to communicate with humans in the human language in like, an interactive voice or like in the in the voice, you know, and it’s kind of it really, you know, it’s, it’s like, it’s not the same as a screen because the screens information is clearly distinct from like, a human eye information, right? Whereas this voice, you know, as soon as you embed the object somewhere, and you don’t see it, well, you know, of course, it’s robotic for the moment, but it gives this you know, immediate human presence, which is not produced by a screen, for example. So it’s quite interesting because from ontological perspective, it’s like the first object that can communicate in the human language. And, and maybe that’s what makes it so interesting in terms of communicating those nonhuman perspectives that basically you, you know, you have this thing that can be like an ambassador of objects, basically, you know, because they can speak to you in English basically, and tell you what it means to be an object. So

Adrian Demleitner 23:18
that’s a that’s a very good point. Do you know this book relating to things? Oh, yeah, yeah. Somebody? I think Claire mentioned to me already, it’s a collection on texts. And there is one from iliza Chuck rd. I think she’s, you know, her work. Not really. I can send it to you. Maybe that’s, that might be interesting. So she is I think she’s in the Netherland. I can send it to you the specific. she, her approach is just to put sensors everywhere. And then like that lets the object speak or let patterns emerge that you can work with?

3j14j0 24:18
Ah, yes, yes, yes. What what was there was a recent project, I think, there was like, was it her now, maybe it was somebody else. But the idea was that here they somehow like they put voice assistant or voice assistant into plants, and used it as like a sort of, it’s not like a I just quite liked more the sort of the, you know, philosophical application was not really that interesting, but it was the, the idea that through the voice assistant, you know, they they put some sensors in the plants to like a measure, you know, moisture, like soil quality and stuff like that. They put like a voice assistant, which means that and they would emulate. So like instead of, you know, just saying, Oh, your plant needs something, the plant will use the human voice to say, Oh, you know, I need this or I need that or like, thank you and stuff. And then the humans, you know, humans could like talk to it and say, so what do you need now the most, you know, in this kind of thing. So it’s like this sort of, again, anthropomorphize, interesting interaction, you know, the way the plants with the help of voice agent. Alright.

Adrian Demleitner 25:31
So I think one last point that that I would love to look at, at with you is that you have the sentence that currently home automation is inherently aimed at optimization. And in another paragraph, you write embedded behavior patterns beyond consumer world oriented reasoning. So it seems like you’re also touching a bit to break this, this descriptive order of things, you know, the clear input and output and not going towards something a bit uncertain, you know, ambiguity, ambiguity, no uncertain. And I was, I was wondering how this, this, this relates to agency, in the nonhuman, you know, that they are not just to our bidding, you know, very scripted that that was the same input comes the same output, but they have a kind of a life on their own.

3j14j0 26:38
Yeah, I mean, there is the problem, especially with these kind of, you know, with domestic appliances is that a, you know, the scope of them is to be produced on mass, and then one object fits all, you know, so you make one Alexa, and then you just put it everywhere. And the other thing is the safety, you know, is that people somehow I say, it’s, like, a very interesting paradigm, you know, it’s like how we sort of this, this consensus again, I’m like, I don’t think that I’m really sort of strong or capable enough to really imagine, you know, how does like, What does, you know, what does it mean for them for the object to evolve from inside, you know, I’m still talking about from, like, my perspective, how I see things, but, but there’s this, like, interesting thing, and let’s say, like, so, let’s, you know, not pretend, but let’s like, embrace the role of a god, you know, when we make an object, right. And it’s interesting how, like, I was the other day to a couple of weeks ago, I was giving this talk about basically how, you know, versus systems are kind of well operating, you know, as a, as a sort of entity of sorts, who should be operating. And then one of the points that I was making is that in showing different videos is that when we imagine, you know, the first thing that you that you do if you’re especially if you’re like a young tech savvy or sort of were taken to gastric bypass, and is that you test their limits, right? You just saw what can Alexa, do you know, how far how many things can the answer, you know, what if I asked questions about love, you know, what if I, and then you know, and then you build this thing that works into either utopia, or dystopia, you know, so, so this is the, we tested and then we tested based of our based on our utopian or dystopian visions of sorts, which like, one, you know, on one side will become your, her, you know, will become this object of love or some sorts, right? And then the other one is like that it will become the wasn’t in house, you know, the dystopia of like, you know, where it refuses to open doors and then like traps you and stuff. And and then so this is like what we humans as gods are kind of imagining, right, and this is like, where our limits are kind of brain links are stretching and we put these devices under this stress on this condition, you know, rather Can you love me or can you trap me in my house and kill me, right? But then at the same time, right, the the scenarios that we develop will never reach these levels because it’s super safe, you know, so we’re building like this really safe technologies and we’re putting all kinds of safe brackets You know, there’s there’s no inherent boundaries in the programming which say that you’re the voices and has to always reply to the same way or that it has to reply every single time right? But there are all these like personality programming, you know, it’s kind of emulates within like a very, very safe way you know, so it will never sort of disobey commands or it will never so it’s it’s interesting, you know, because we like if it this evade, we wouldn’t be genuinely impressed by the quality of something that we built right? That’s so true. So that’s quite from the human perspective. I think that’s like, that’s quite fascinating to look at, you know, the thing that’s sort of how he had the vision versus how we actually build it. And especially in domestic appliances, you know, because you probably can look at more experimental things, but I’m sort of more interested in maybe in those offerings, because in the end, that’s what kind of shapes the human you know, perception towards technology in the end, like, our phone is what you know, listens to us and stuff. And the Roomba is what like replaces the the traditional house, or something.

Adrian Demleitner 30:42
I also found this one study called “My Roomba is Rambo”. And it’s really one of my favorite studies ever, because it, it goes into why roombas are very relatable home appliances.

3j14j0 30:58
I think maybe you sent it, but I didn’t. I didn’t have tested it wasn’t meant to link already. No. Okay. Yeah, that’s, that’s good. Yeah, so our, our current one, like our second case study is the Roomba, it’s like, it’s, it’s quite funny, because we, again, like the, the idea behind the Roomba is that the voices system is actually quite capable, you know, device, in a sense, compared to all the others, it’s sort of hierarchically already on top of the chain. Because not only it can, you know, talk to humans, but it can also govern over all the other devices. So it can turn things on off, you know, it can sort of activate deactivate stuff, you know, it can go to the internet and, like, look things up and decide which information is relevant and which is not as useless. So, whereas a woman is a bit stupid, it just like, bumps into the corners, even like the new ones, they they optimize their trajectories, but they you know, that’s pretty much it, you know, in the end, you just said best. So, so we’re kind of saying that it’s interesting, because when you look at the, you know, at the, from a device perspective, when you look at the home ecology, the the interesting part is that, well ruma sort of becomes the legs of understanding the home space in physical terms. If the voice assistant is a bit, I am who you ask me, you know, I am what you asked me and and then through that I understand what is the purpose of like a voice assistant, you know, how do I serve the home by being diverse, then the Roomba is actually you know, it’s its trajectory? it’s maddening it’s like seeing where the objects are where the cats are, you know, what kind of home Do you live you know, again and and all that. So we like we’re currently using get this this spiritual guru, basically, because we said, you know, there’s always, if you have a tandem of two objects are like of two things. It’s like that the pinky in the brain was that that were like, one is that this evil mastermind and the other one is like, this is a bit silly, but spiritual, you know, animals. So So the idea is that one the voices and this kind of this knowledge gathering, you know, entity that is trying to understand the concept of self Roomba is this like, spiritual randomizer, where, based on where the bumps You know, it didn’t kind of messes with the brain of devices that basically, I could approach very much. Okay, so so that’s, that’s it? Yeah, the project is gonna say, honestly, that the room is finished actually, that because the room was completely done. Now, the The only thing is that the voices system because it’s like a bit of a complex things, it’s still we’re kind of finishing it, but like, constantly running into one bug after another, but but we’re still I think, in September, it’s gonna, it’s gonna come out properly and stuff. Okay. I’m looking forward to that. Yeah. That was a while in the making. No. Yes. Yes. Yeah, let me know the project, actually, it’s quite funny, we realize that all of my projects will take a long time. And then it’s funny because people like I don’t think that afterwards now, especially when you put it in the gallery place. I don’t think that people really have time to dive so much into the detail, you know, sexually, things that are built quicker and communicate and like a more simple way are probably more effective for for places like a gallery, but it’s another scenario when you actually put it in the home and you get to live with it for a while because then it’s kind of you know, then becomes this weird you know, inhabit like it becomes part of the home you know, and it’s kind of funny to have an object like this as opposed to just having like an Alexa you know that anyways, whatever. Yeah.

Adrian Demleitner 35:01
Yeah, cool. I think I went through the questions I had and and I get much more. You said some really cool things. I gathered some really good insights to continue.

3j14j0 35:18
Yeah, if I if I run into any references and stuff, I will also let you know. I told you like there was the study by I don’t know, I was looking for it again. And I don’t see where he was not crazy, like interesting. Yeah, but it was about elderly people. Basically it had it and yeah, yeah, like that. Yeah.

Adrian Demleitner 35:37
There’s a lot of stuff going going on in academia as well around voices systems in the humanities as well. And yeah. Yeah, that’s a puzzle. There’s something good comes out. I hope so. So I would say I’ve got four minutes left. I thank you very much for for the time. I will stop recording here.

Transcribed by https://otter.ai