Friendly Computers
Why are computer interfaces always so friendly?
For more information on Aleph Insights visit our website https://alephinsights.com or to get in touch about our podcast email podcast@alephinsights.com
Transcript
Hello and welcome to the Cognitive Engineering Podcast produced by me, Fraser McGruer, for Aleph Insights. In this series of podcasts we take a look at interesting topics and discuss what we think they tell us about analysis and decision making. I'm here with Chris Wragg, Nick Hare and Peter Coghill of Aleph Insights and this week we're talking about why computers are so bloody friendly. So, yeah, the other week I was doing something on my computer, as one does.
Speaker B:Browsing porn.
Speaker A:And, um... Didn't deny it. So you completely threw me off track.
Speaker B:So I can just interject.
Speaker C:Don't think about the porn and get on with the podcast.
Speaker B:No, but I just, there was a moment in parliament about 30 years ago when Michael Heseltine stood up to deliver a speech and he had his hand in his pocket and Dennis Skinner shouted, stop playing with yourself. And Heseltine later said, well, the problem with that is if he left his hand there, it would be like he was still playing with himself. But he took his hand out.
Speaker A:It would be an admittance.
Speaker B:I think you're in the same position now. Yeah. But carry on. Please carry on.
Speaker A:Thank you. So I think I was starting up my computer. He wanted me to install something. And I forget what it was, but it was just so damn happy about it. And so congratulatory about the whole thing. And, um...
Speaker B:We're excited that you're choosing to use this upgrade.
Speaker A:Yeah. And even, I think Jack D once said, even when you turn a computer on, it goes, And, you know, his, his sort of thing. God, what's the big deal? What are you so happy about?
Speaker C:And you fulfilled your minimum function, which is turning on. I think was what he said.
Speaker A:Yeah. Yeah. Um, and it occurred to me that, hold on, we've, we've kind of got used to computers being so friendly with us and so enthusiastic and, uh, with us that we've kind of might have forgotten, you know, or we've become inured to it. It got me thinking that the reason why this kind of, this sort of way of talking, this way of communication has become, you're so ubiquitous and so successful is precisely because they're like that. So let's say that the Soviet union had developed the software that we all start, that we will use all the time. That actually, that, that was an impossibility that could never have happened.
Speaker B:In Soviet Russia, computer programs you.
Speaker A:Yes. Right. I'm not making my point very clearly. Do you get what I'm trying to say?
Speaker B:I totally get what you're saying because I'm as infuriated as you are. Right. And I think this boils down to this, this issue of the computers trying to be your best mate. Yeah. And for me, the rot set in, uh, in windows 95, right? Windows three windows 3.1. Uh, it was pretty work a day. I mean, it, you know, it was, uh, it was fairly functional, you know, it was nice graphic. It certainly was revolutionary at the time, unless you're a Mac user, in which case you'll claim that it was all ripped off from, from Mac stuff. But you know, if you wanted to find, oh, that's perfect. Is that Alexa? Do we want to just get interrupted by Alexa? We've been interrupted by Alexa. That is absolutely perfect timing. So, um, and then windows 95 came along and, and I think they'd realized that now their target market was no longer businesses. And, you know, windows 3.1, you had, if you wanted to find your files, you'd navigate, you'd go through file explorer and you'd get to, you know, your C drive and it would be called the C drive. And then, you know, you get to windows 95 and suddenly there's a thing called my computer, my computer. And that drove me mad. I was so annoyed by that because it just seemed illogical to have an icon on my computer called my computer. My computer is sitting in front of me. That's not my computer. What that is, is a set of links to hard drives and files. It's not, it's not the same thing as my computer. And I, and I feel like ever since then operating systems, you know, apps, mobile phone app, everything has just been, it wants to be your best mate. It's always like talking to you when you install it, you know, hey, would you like some help with that? Would you like, you know, would you like to get on our mailing list? Would you like to know about these other apps? It's all of that stuff. It drives me nuts.
Speaker A:Yeah. So, I want to hear from Peter. What do you reckon?
Speaker C:Well, the sort of Californian positive spirit kind of attitude to a lot of these AIs probably comes from the fact that they come from the global tech cradle in California. Silicon Valley.
Speaker A:I mean, I think, you know, we're all sort of fairly sort of miserable English types or British types, right?
Speaker B:Yeah, I'd like to turn the computer on and for it to say, what can I do you for?
Speaker C:Well, yeah. We're not wanting to invoke too many kind of national stereotypes. Yes. A British designed AI you'd expect to be slightly sarcastic and wacky.
Speaker B:You've jumped straight to AIs. Well, we're talking about interface design.
Speaker C:But yeah. But it's an interesting thought. Sarcastic and wacky. So, you might just hide your my computer icon just for a laugh. Just on a random basis. And you might expect a Russian designed AI to be a bit sort of morose and dark about things. It wouldn't.
Speaker B:It just wouldn't. Wherever it was, it wouldn't be where you wanted it to be. Yeah. Computer is not work today.
Speaker D:I mean, presumably, though, you know, from the AI perspective, you know, they start out friendly in California because that's where they were programmed. But surely the whole, you know, the future that's ahead of us is ones that adapt to us as individuals and learn whether we're cynical, friendly, you know, what we like in the morning and what we like in the evening.
Speaker B:I feel like we've run a very long way ahead very quickly here. Like we're almost about to do a touchdown.
Speaker A:So bring us back then.
Speaker B:Well, so, well, I will. And I think it's to do with how humans, let's start with the question of how humans want to interface with machines and the extent to which that's something fundamental about our attitude to them because the machines or whether it's something whether it's whether it's a question of the designers of the machines wanting to trick us into thinking of something else. So let's just start with that. This concept of how we interface with machines because, of course, that's changed dramatically over quite a short space of time. In 30 years, we've gone from you needing to be reasonably technical and to speak the machine's language to be able to even just to use something as high level as something like DOS. You kind of needed, I mean, you certainly couldn't type, you know, you couldn't talk to DOS. You had to use DOS's language. It's kind of restricted vocabulary to be able to do things that we would think of as very simple today. You know, gradually towards being more and more user friendly and more and more like a real life, more like a thing, more like a kind of other person who is doing stuff for you, more like a kind of servant of some kind, a kind of, you know, helpful person in your house, like the kind of robots you used to get in sci-fi films. So we've moved away from it being we really had to go all the way to the computer to understand how to get it to do stuff to it trying to do more and more for us. Um, now, is there a point at which it needs to stop and be just to be a computer and act like a computer? Or is there, do we really want to go all the way down the Siri, Alexa, Cortana kind of route where, um, you know, they are, we treat them like people and we talk to them like people.
Speaker A:Okay. Peter?
Speaker C:Well, I think they, they probably just responding to what they perceive as being the prevalent demand. The designers of the programs, not the machines themselves. There are different, there are sort of different styles. So, um, Alexa, I think is quite, uh, she's, she's quite sort of business-like but friendly without being officious in the way she talks. Or supercilious. Yeah. She's sort of helpful, but not kind of chummy. So she, yeah, so that, that, that suits my kind of style human in human ish, but not trying to be too human. Just like Peter himself. Just like me. Whereas you get lots of sort of chatbots on websites. If you go to a big software website, there'll often be a thing bottom right that is kind of hi there. If you've got any questions, just ask. And you could just have a box saying questions here. And that would do, that would be, that would get the same message and be much more less messy in the way it's written.
Speaker D:But, but I, I think something Peter said there was a really interesting tell, uh, and why they are doing this, right? He referred to Alexa as she, uh, and he's already given it a gender and personified it. And that's the whole point, right? That, you know, it's, it's this, it stems from this, uh, or at least it partly stems from this, uh, whole, you know, marketing approach, which is anything you sell, you're trying to generate a, you know, a personal rapport with somebody in order to sell it to them and to get them to buy into the goods. It's why, you know, every, every craft beer or, or packet of organic crisps has a, you know, a whole spiel about the people who made it in their, their little farm and how, how nice they are. Uh, and this is just an extension of that. It's the, the good itself takes on, uh, uh, you know, is kind of anthropomorphized to some extent. And, uh, we strike up a relationship with it and, and that buys sort of brand loyalty
Speaker B:or, you know, I think, I think you've nailed it, right? Because so now I'm going to go to the NF cupboard of useful concepts and bring out an old friend Fisk's relational models theory, which is, I think we talked about it before about the way that we interact with companies, the way that companies want us to interact with them, right? So this is the idea that there are basically four kinds of relationship you can have when you're, um, well, when you're kind of sharing resources or at least distributing resources. And of course, when you're doing a job together, which you and the computer are, it's not dissimilar. And he says there's, he's observed that there's communal sharing where you're all kind of, um, you know, just parceling out stuff, uh, equality matching where you're trying to make sure that no one's getting, uh, getting stiffed. Uh, like if you're, you know, splitting a bill, authority ranking where you're the boss and you decide who gets what and then market pricing where it's totally, uh, you know, dehumanized and it's a single price and everyone gets the same, uh, everyone pays the same price. And I, and I feel like here, this is computers, certainly how I perceive it, why I get irritated with it because I feel like they're misjudging the relationship I have with them. As far as I'm concerned, I am in an authority ranking relationship with the computer. And in case anyone's wondering, if there's any computers listening, I'm the boss in that relationship. I decide what they do. They're not my friend, right? It is as inappropriate as a employee who, you know, is getting chummy with you and wants to talk about goss and private life and stuff. It's just, it is irksome and irritating for that reason. And I think that's, I think that's more or less it. I think Chris has nailed it.
Speaker C:Um, that brings us back nicely. Are these, these, these interfaces, which are a product of the Silicon Valley kind of style of organization, which are a bit more like that. They're a bit more communal and chummy. They pretend to be, yeah. They pretend to be, or they try to be, or they aspire to be. Whereas the traditional kind of Western European corporate culture is very much hierarchical and feudalistic in terms of there are these overlords. And then you are the underling who does work based on their, on their, on their whim.
Speaker A:And that goes back to my original question. Isn't it precisely that kind of organization that's more informal, that's more horizontal in its structure, that's got the greater chances of success in producing, um, innovative technology. So it's all, it's inevitable that we end up with these really friendly computer interfaces.
Speaker C:I don't, I don't think that's true at all. I think it's possible to nurture true innovation in a very sort of hierarchical traditional company. Plenty of companies do that. I mean, um, and yet the reality though is that all computer interfaces are really friendly.
Speaker B:I think you're making a bit too much of this relationship between, look, it's not, it's like advertising talk. Advertising talk also wants to be friendly, right? Again. And I think it's this mismatch between, because I think in the previous podcast, I mentioned this in the context of restaurants want to make you feel like you're there, you're, you're in a kind of family relationship and feel comfortable like you do if you're having a meal with your family, but it's actually a market pricing relationship. So they try and disguise the fact that you're basically there paying for them to do stuff. They want to make it look like they're kind of giving you stuff for fun. And I, and I think, you know, the language of advertising, again, they want to get you out of a market pricing relationship and make you feel like you're in a kind of communal sharing relationship with them because that's where we're most comfortable. And I just think it's, it's sort of, it's not necessarily Silicon Valley. It might just be because that's just a good trick.
Speaker A:It works. Anyway, just one thing I want to come back to is, there was an, there was an article yesterday, just yesterday. I don't know if you saw it. I think it's from Google, which was, they played some examples of an AI ordering, um, first of all, making a point, an appointment for, for a haircut for someone. And then another one of the AI, so this is over the telephone. Cool. Interacting with someone to book a restaurant, um, table. Did any of you see this? No, but go on.
Speaker C:I definitely need both of those things. Okay.
Speaker A:And what was, and so there were a couple of things that were interesting. And so essentially, like I said, I think it was Google. And a couple of things came out is they put in, when the AI was speaking, they put a lot of, uh, uh-huh, uh-huh, that kind of thing. And that worked really nicely. And, and the, and often the people, or certainly in the examples they showed, the people they were dealing with didn't realize it was an AI.
Speaker B:I need some more information. Were the people they were dealing with aware that they were talking to an AI? Okay.
Speaker C:I suppose if you're trying to mimic human, a human, then yes, you need to introduce the verbal tics.
Speaker A:But the difficult, yeah. But the difficulty is getting those tics in the right place and doing all the programming behind that. But also when, so in the second example was really interesting where it turned out you didn't need to book at this restaurant if you had a certain number of people and the AI got it, the AI understood. So that's the first thing. But the second thing was, I think something Peter said earlier on, or was, was what, or maybe it was yourself, Nick, what we want from our AI and how friendly do we want them to be? Because one of the things that, again, let's say it was Google said was the, but actually when the day comes, when we actually release this kind of concierge service, it will always be the case that we'll let the people know that actually it's an AI they're talking with.
Speaker D:Well, so, so I think this is, this is coming even quicker than, than that because you know, Google are in the process of launching their smart compose service for, for emails, which you've probably seen it to some extent already when you get an email, it's guessing what your response might be. And you can simply, it's now making that more sophisticated. And presumably that is going to learn more about your style of communication, which of those options you pick, it will offer more, more often. And I think this type of thing gets to the heart of why we find friendliness in computers, particularly difficult to deal with. It's, it's the presumptuousness of them pretending to be human. And we know they're going to supplant us at some, some stage. And we, we have a deep Freudian, deep Freudian resentment of, of that. And this, this, you know, this, this fact that now you're going to have. We're going to let him get away with that, but carry on. Now you're going to have this dilemma of other people writing emails back to you, and you don't know whether those have been composed by themselves, part composed by themselves, or they've just clicked, whatever the response will happen to be. How will you then feel about those responses coming, coming back to you? And it's like getting a handwritten letter that you find was actually not a, not a genuine handwritten letter.
Speaker C:Well, there's a big industry and people writing handwritten letters. Exactly.
Speaker B:In that fantastic film, which everyone ought to watch, Her. Yeah, that's his job. He does, he writes handwritten letters.
Speaker D:But imagine if it wasn't him. I mean, at least, at least that's a person, you know, when you were talking about.
Speaker C:But I see this as just an extension of the spellcheck function in, in your browser or, or your word processor. I'm totally reliant on that being quite dyslexic. I don't do words very well. And I'm utterly reliant on the spellcheck in terms of putting meaningful sentences together. And even then, as Chris and Nick will attest, I get it wrong quite frequently. But that to me is just a, is a really useful crutch slash coping mechanism for something that I just not able to do.
Speaker D:I mean, the interesting loop will come when, of course, you know, you, you send back the Google Smart Compose to somebody who then turns another one back around to you, at which point we all presumably go home and put our feet up. Yeah, although, you know, if I got an email.
Speaker B:That's ideal, isn't it? But if I got an email from Peter with no spelling mistakes, I think a computer's This is really unconvincing.
Speaker A:Okay, so where do we go? We're close to wrapping up, but not quite there. Anything anyone want to say?
Speaker C:We could, we can, we've been talking a lot about the kind of human interaction in AI. We could extend it a bit further to general interfaces, as Nick was trying to go with, with the My Computer icon that you hate so much. You can rename it if you want.
Speaker B:Yeah, I did. That's the first thing I did. I called it something like Gerald.
Speaker C: computer related interfaces,: Speaker B:Yes. Like, you know, a Stephen Hawking voice or something. I have to say when I'm interacting, even if it's something like, you know, there's voice activated menus when you phone a call center and you get a voice activated menu where it says, tell me what you want. And you're supposed to say, you know, complaints or something. I know the idea is that they're supposed to be humanizing the experience, but what I experienced as me being dehumanized. I am having to, you know, a bit perform like a performing monkey for a computer at the other end. It dehumanizes me. And so I think, you know, yeah, user-friendly basically doesn't need to mean actually friendly.
Speaker C: l. In the same way that early: Speaker D:People have got a bunch of new toys and they're playing with them.
Speaker B:The application designers were so interested in whether they could that they forgot about whether they should.
Speaker A:You stole my line, Jeff Goldblum from Jurassic Park. I was about to say exactly that. Before we finish, just talking nothing to computers, but relationships and interfaces being as they should, not, you know, faking it. A pet hate of mine, OK, is people calling me mate, right? When I don't think they fall into a category of person who should be calling me mate.
Speaker B:If they call you mate, they're probably about to start a fight with you.
Speaker A:Well, OK, so I actually wouldn't mind that.
Speaker C:I don't mind that. In Scotland, they get around this. They've got a word called pal, you know, you call people pal, which is like someone you're trying to be friendly with. You don't know their name, but not mate. Mate is reserved for people who genuinely know you. Now pal is also fighting talk. Yeah, come on, pal. You're going to the wrong pubs, mate.
Speaker A:So anyway, people who I like calling me mate are my mates. Yeah, but we don't.
Speaker B:I've never, never called you mate in my life.
Speaker A:Yeah, no, you haven't, no. But if someone is my friend, I don't mind them calling me mate. I quite like it. If someone, if I don't know them at all, just some bloke in the street calls me mate, I quite like that. That's fine. That's friendly, right? But what I don't like is someone who vaguely knows me, maybe in a professional sense or something. An acquaintance.
Speaker D:Is it younger people calling you mate you have a problem with?
Speaker C:You'd rather they called you Mr Fraser. Would they call, do you like if they call you bro?
Speaker A:I definitely wouldn't like that. What about geese? No. Yeah, I once got called geese-a-fella about 10 years ago. That's an interesting one.
Speaker B:But I think this is, this goes back to that misjudging relationship. And that's, I think what this all boils down to. Yeah, exactly.
Speaker A:Okay. Cheers, mate. So we'll stop there. Right, bruv. Thanks as always for listening to the Cognitive Engineering Podcast. I'm Fraser McGruer. We've been here with Nick Hare, Peter Coghill and Chris Wragg of Aleph Insights. Until next time. Bye-bye.
