WEBVTT - Poker Face 0:00:04.760 --> 0:00:08.480 Sleepwalkers is a production of I Heart Radio and unusual productions. 0:00:11.280 --> 0:00:14.040 We can choose to have a poker face, but the 0:00:14.080 --> 0:00:17.560 point is that our bodies are still reacting and what's 0:00:17.680 --> 0:00:22.040 changed is the ability to see those signals. That's Poppy Crumb, 0:00:22.160 --> 0:00:26.040 chief scientists at Dolby Labs and a professor at Stanford University. 0:00:26.520 --> 0:00:30.000 Her work is at the forefront of neuroscience and data science, 0:00:30.120 --> 0:00:33.960 and it's bad news for the poker face pret Solo 0:00:34.040 --> 0:00:37.320 just want to an Academy award. And the filmmakers were 0:00:37.760 --> 0:00:41.520 at my company doing a screening and had we had 0:00:41.640 --> 0:00:45.200 captured carbon dioxide of the audience with her approval, of course, 0:00:45.560 --> 0:00:47.840 and I wasn't actually at the screening. I just saw 0:00:47.880 --> 0:00:50.800 the Ceoto capture and I knew exactly where the climbs were, 0:00:50.840 --> 0:00:53.920 where he abandoned his climbs. As the audience watched Annlex 0:00:53.960 --> 0:00:57.640 Honald attempted to climb El Capitan, their bodies responded to 0:00:57.680 --> 0:01:01.200 the suspense, their breathing change, and thanks to the carbon 0:01:01.240 --> 0:01:04.319 dioxide sensors in her theater, Poppy had a map of 0:01:04.360 --> 0:01:09.120 the audiences emotionally experience. It's this power of the audiences 0:01:09.200 --> 0:01:12.959 on that journey and experiencing it with the filmmakers, and 0:01:13.560 --> 0:01:17.960 it's pretty exciting to see that engagement in the theater 0:01:18.080 --> 0:01:21.679 and have that history. But our breath isn't our only 0:01:21.760 --> 0:01:25.360 tell There's an increasing number of ways machines are becoming 0:01:25.400 --> 0:01:28.840 able to read us, even how hard we're thinking. In 0:01:28.959 --> 0:01:32.480 thermal cameras, you can track, you can look at dynamics 0:01:32.560 --> 0:01:36.680 of blood flow to know stress levels and engagement. Just 0:01:36.800 --> 0:01:40.319 in the infrared signatures, you can understand cognitive load. You 0:01:40.360 --> 0:01:43.720 can then look at micro expressions of facial recognition to 0:01:44.200 --> 0:01:47.120 get past not just if I'm feigning emotion, but really 0:01:47.120 --> 0:01:49.920 the authenticity of what I'm experiencing. That gives us a 0:01:49.960 --> 0:01:52.200 lot of insight about how hard my brain is working 0:01:52.720 --> 0:01:55.720 and how engaged I am. We haven't changed this humans. 0:01:55.800 --> 0:01:59.320 What's changed this ubiquity of sensors and the capacity of sensors. 0:01:59.320 --> 0:02:03.000 The the cost just Fitteen years ago, the cost of 0:02:03.040 --> 0:02:07.600 a typical device would be about maybe. Now you're looking 0:02:07.720 --> 0:02:10.280 at those devices not even having to be close up 0:02:10.400 --> 0:02:14.280 for pennies dollars integrated into every pair of smart glasses. 0:02:14.400 --> 0:02:19.680 Going forward, we're on the cusp of two explosions. The 0:02:19.760 --> 0:02:22.919 power of machine learning to find patterns and make predictions, 0:02:23.280 --> 0:02:28.360 and simultaneously the miniaturization and affordability of cameras and other senses. 0:02:29.160 --> 0:02:33.200 Last episode we talked about facial recognition and surveillance by governments. 0:02:33.560 --> 0:02:37.160 But when machines contract how we're feeling, our most private 0:02:37.240 --> 0:02:41.400 selves become readable, and while that may sound frightening, it 0:02:41.480 --> 0:02:45.120 also holds enormous promise for many parts of life. From 0:02:45.200 --> 0:02:49.200 beginning to end this episode, we look at what's changing 0:02:49.400 --> 0:03:06.320 and what's possible. I'm as Veloshen. This is Sleepwalkers, my 0:03:06.480 --> 0:03:11.440 Ma Mama. So there are quite a lot of situations 0:03:11.520 --> 0:03:14.320 where personally, I don't actually want to be read. I'm 0:03:14.320 --> 0:03:17.200 not sure about you, like I want to hold them 0:03:17.280 --> 0:03:20.200 like they do in Texas play, Like when I'm playing Texas, 0:03:20.200 --> 0:03:22.600 hold them on people to know what I'm thinking, right, 0:03:22.680 --> 0:03:24.880 and not just at the poker table. In fact, our 0:03:24.880 --> 0:03:27.200 society kind of relies on the idea that we can 0:03:27.560 --> 0:03:31.040 look one way but feel another. Obviously, in plays like Hamlet, 0:03:31.200 --> 0:03:36.240 interiority is dramatized, but but more broadly, society is where 0:03:36.240 --> 0:03:38.480 people have no privacy tend to be a bit scary. 0:03:38.600 --> 0:03:41.080 It is scary because the last thing we have on 0:03:41.200 --> 0:03:43.560 earth is our privacy. You know, it's like people have 0:03:43.600 --> 0:03:45.920 this impulse to share everything on Instagram and give away 0:03:45.920 --> 0:03:49.040 their name to a company that wants to sell them jewelry, 0:03:49.400 --> 0:03:53.040 and it's just like, truly, our deep our secrets are 0:03:53.040 --> 0:03:56.440 the last thing we have, well the last thing we had. 0:03:56.760 --> 0:03:58.720 Right probably did a full ted talk on this, and 0:03:58.760 --> 0:04:00.760 the thing that I took away from it is will 0:04:00.800 --> 0:04:02.800 we live in a near future where a slasher film 0:04:02.840 --> 0:04:05.880 will be edited with people's biometric data in mind? Poppy 0:04:05.880 --> 0:04:08.160 suddenly sees that on the horizon, and she has a 0:04:08.280 --> 0:04:12.440 term for technology that starts to understand us. Empathetic technology 0:04:12.480 --> 0:04:15.120 is the idea that you know, it's it's not technology 0:04:15.120 --> 0:04:18.159 that empathizes with me or technology that is trying to 0:04:18.200 --> 0:04:21.359 emulate human empathy. It's technology that makes use of my 0:04:21.440 --> 0:04:24.880 internal experience to be able to integrate that as part 0:04:24.880 --> 0:04:28.680 of its interface. Today, it's impressive that Poppy can understand 0:04:28.720 --> 0:04:32.359 an audience's emotional journey watching Free Solo by tracking the 0:04:32.440 --> 0:04:35.200 levels of CEO two in their breath. But tomorrow it 0:04:35.240 --> 0:04:38.360 could lead to new kinds of art and entertainment that 0:04:38.440 --> 0:04:42.520 responds to us. Really great hip hop producer I was 0:04:42.600 --> 0:04:46.080 talking to wants to create music that is personalized, almost 0:04:46.080 --> 0:04:48.520 like a tailored suit for individuals. So you start to 0:04:48.520 --> 0:04:52.800 think about a very dynamic integration of the human experience. 0:04:53.120 --> 0:04:56.279 That experience becomes something that our technology can be aware 0:04:56.279 --> 0:04:59.480 of and optimized for. I want my technology to make 0:04:59.520 --> 0:05:01.719 the right di visions so that the experience I have 0:05:01.839 --> 0:05:06.480 with it is seamless. Seamless such a seductive word they 0:05:06.560 --> 0:05:10.160 named a food delivery service after it, but a dangerous 0:05:10.200 --> 0:05:13.720 word too, because for technology to read and respond to 0:05:13.800 --> 0:05:16.880 us in real time, it needs to make decisions about 0:05:16.960 --> 0:05:20.080 us on its own. You may remember last episode we 0:05:20.120 --> 0:05:22.880 spoke with Lisa talia Moretti about some of the risks 0:05:22.880 --> 0:05:26.400 of facial recognition technology, but that's not her only area 0:05:26.400 --> 0:05:30.279 of research, something that I was looking into really recently, 0:05:30.560 --> 0:05:36.120 as our relationship with technology is completely shifting. So we're 0:05:36.160 --> 0:05:39.520 moving from a relationship with technology where we are asking 0:05:39.520 --> 0:05:41.960 it to do something, you know, it's a pure sort 0:05:42.000 --> 0:05:46.160 of input output, And what now we're moving towards is 0:05:46.200 --> 0:05:50.680 a relationship with technology where we are trusting technology to 0:05:50.760 --> 0:05:55.080 make decisions on our behalf. Lisa teaches at Goldsmiths in 0:05:55.160 --> 0:05:58.760 London and Cardiff University. She told us about how her 0:05:58.800 --> 0:06:02.640 students have encountered chnology making its own decisions about their 0:06:02.680 --> 0:06:05.760 future prospects. One of the things that the students are 0:06:05.800 --> 0:06:10.120 starting to do is to game the algorithms that are 0:06:10.160 --> 0:06:15.680 being used to mind through candidates cvs. And so what 0:06:15.720 --> 0:06:18.720 they figured out is if they put right in white 0:06:18.760 --> 0:06:24.560 text anywhere on their CV UM Cambridge, Harvard, Oxford, they're 0:06:24.560 --> 0:06:27.279 more likely to get through to the interview process. Does 0:06:27.360 --> 0:06:32.080 students know that recruiting algorithms prioritize applications from certain schools, 0:06:32.360 --> 0:06:34.880 so they pepper their applications with words they know the 0:06:34.960 --> 0:06:38.119 algorithm will like, but written in white. So the human 0:06:38.160 --> 0:06:41.800 recruiters are not the wiser. They're marketing themselves straight to 0:06:41.839 --> 0:06:46.360 the AHI. They're gaming the algorithm system, which I think 0:06:46.400 --> 0:06:51.479 is pretty genius. There's also certain things where students or 0:06:51.600 --> 0:06:55.159 candidates who are having to conduct their first interview in 0:06:55.200 --> 0:06:58.359 some companies purely online and there's no person on the 0:06:58.360 --> 0:07:02.480 other side. You're essentially talking into your webcam, and there's 0:07:02.520 --> 0:07:08.600 algorithmic technology that is recording your voice, that's listening for intonation, 0:07:08.880 --> 0:07:11.440 that's listening for the types of words that you're saying, 0:07:11.520 --> 0:07:14.640 like if you use smart words or your language isn't 0:07:14.720 --> 0:07:17.360 perhaps at a level that they would think is appropriate 0:07:17.360 --> 0:07:22.280 for business. But do we want computers to deny opportunities 0:07:22.280 --> 0:07:25.240 to job applicants that may be qualified but not fully 0:07:25.280 --> 0:07:29.720 polished without human review. And the algorithms weren't only analyzing 0:07:29.720 --> 0:07:33.560 the student's words. They're also looking at your facial features, 0:07:33.600 --> 0:07:36.600 and so they can say if you're nervous or shy. 0:07:36.760 --> 0:07:39.440 And some of my students have said that if they 0:07:39.760 --> 0:07:43.040 very quickly use like hand gestures, they confuse the camera 0:07:43.080 --> 0:07:45.520 and the camera councy if they were nervous or shy 0:07:45.720 --> 0:07:49.320 for those particular moments. There's something quite hopeful about these 0:07:49.360 --> 0:07:52.760 students averting the algorithms designed to read them. It's not 0:07:52.800 --> 0:07:54.680 quite the Summer of six y eight, Cara, but the 0:07:54.720 --> 0:07:57.320 youth have still got something. Well, it's true that you 0:07:57.400 --> 0:07:59.840 and I both look very good in black leather and 0:08:00.160 --> 0:08:05.800 are considered cyberpunks. That's how we met that cyberpunk rally. 0:08:05.880 --> 0:08:10.560 I was reading about this Kickstarter campaign called reflectacles like specticles, 0:08:10.560 --> 0:08:15.000 but reflective, that's right, And that's because they reflect invisible 0:08:15.040 --> 0:08:17.960 and infrared light. When you look back into a camera 0:08:18.040 --> 0:08:20.640 that's watching you, which is like an ultimate that's like 0:08:20.680 --> 0:08:23.480 a techy middle finger, like I'm gonna look right back 0:08:23.480 --> 0:08:25.960 in this camera. It's just gonna like buzz light back 0:08:26.000 --> 0:08:29.520 at it completely. So you know, there are methods obviously 0:08:29.720 --> 0:08:33.720 for resistance, But what I'm worried about is that algorithms 0:08:33.800 --> 0:08:37.160 are very smart and they will wise up and be 0:08:37.240 --> 0:08:39.240 harder to trick. Right, and showing up for your job 0:08:39.240 --> 0:08:42.720 interview and reflecticles probably carried his own burden as well. 0:08:42.840 --> 0:08:45.040 It is how I got this job, though, But we 0:08:45.040 --> 0:08:48.640 should also ask ourselves why these companies are using AI 0:08:48.679 --> 0:08:52.280 to filter candidates and conduct interviews, And of course it's 0:08:52.320 --> 0:08:55.839 really about saving money and saving resources, which brings out 0:08:55.840 --> 0:08:58.240 the big question of the series. Who benefits from this 0:08:58.320 --> 0:09:03.240 new technology Amazon and Facebook and Google. When we come back, 0:09:03.440 --> 0:09:05.360 we look at the economics of giving up our data 0:09:05.559 --> 0:09:15.319 and what we get in return. Sensors and AI to 0:09:15.400 --> 0:09:18.040 analyze our response to movies or decide if we're a 0:09:18.080 --> 0:09:20.600 good fit for a job. May sound like the stuff 0:09:20.679 --> 0:09:25.319 of dystopian science fiction, and that's because it is set 0:09:25.400 --> 0:09:29.400 up a progner and Tilmon Rope. I'm placing you under 0:09:29.480 --> 0:09:31.199 arrest for the future murder, Sarah Marks, you have a 0:09:31.200 --> 0:09:35.640 man his head. Yeah, oh, gosh, Well, it's interesting you 0:09:35.679 --> 0:09:42.120 mentioned Minority Report because, um to this day, so many 0:09:42.200 --> 0:09:45.400 years later, decades later, I'll be in some meeting in 0:09:45.440 --> 0:09:48.520 Silicon Valley and We'll be looking at some gadget and 0:09:49.600 --> 0:09:51.520 so I say, Wow, this gadget is great. It's like 0:09:51.559 --> 0:09:54.720 from Minority Report. It's so cool, And I'm like, that 0:09:54.840 --> 0:09:58.120 was supposed to be cautionary. That was a description of 0:09:58.160 --> 0:10:00.480 the bad world. That was what we want to avo wide. Oh, 0:10:00.480 --> 0:10:04.360 for God's sake, that's Jaron Lania. He's a research scientist 0:10:04.440 --> 0:10:06.920 at Microsoft and in the eighties he coined the term 0:10:07.160 --> 0:10:11.160 virtual reality after helping invent the field. Jaren's thought a 0:10:11.160 --> 0:10:13.839 lot about what our relationships with technology mean for us, 0:10:14.120 --> 0:10:17.560 So when Steven Spielberg was making Minority Report, he called 0:10:17.600 --> 0:10:20.760 on Jarren to act as a technology consultant. Mostly, what 0:10:20.840 --> 0:10:24.520 I've taken from Minority Report is that just trying to 0:10:24.559 --> 0:10:30.800 do cautionary portrayals of technology actually backfires, because there's some 0:10:30.840 --> 0:10:33.480 way that it's a little bit like when you show 0:10:33.520 --> 0:10:36.240 the Life of Billionaires, people don't get angry about like 0:10:36.240 --> 0:10:38.400 why do those people monopol or where they own whole 0:10:38.400 --> 0:10:41.520 islands or something. Instead they say, oh, I identify with 0:10:41.520 --> 0:10:43.600 that person. Maybe I could own a whole island someday. 0:10:43.920 --> 0:10:47.959 And despite our fascination with dystopian fiction, we also have 0:10:48.000 --> 0:10:51.680 a tendency to fantasize about ourselves as the beneficiaries, not 0:10:51.800 --> 0:10:54.960 the victims, of the systems we create, and we tend 0:10:55.000 --> 0:10:58.800 to ascribe those systems their own will, even though we've 0:10:58.800 --> 0:11:02.360 made them. Early in the history of capitalism, Adam Smith 0:11:02.800 --> 0:11:06.199 suggested that capitalism or markets were an invisible hand, as 0:11:06.200 --> 0:11:08.720 sort of a life form. And in the same way 0:11:08.760 --> 0:11:12.200 that you can interpret a market as being this living 0:11:12.320 --> 0:11:15.280 thing just because it's a little beyond our understanding, it's 0:11:15.280 --> 0:11:19.119 a little too complicated to fully predict and fully understand, 0:11:19.240 --> 0:11:21.920 and and that's actually its power. In the same way, 0:11:21.920 --> 0:11:24.880 big computational systems can be a little out of control, 0:11:24.920 --> 0:11:26.760 not entirely, but even if they're only a little bit, 0:11:26.800 --> 0:11:30.040 you can interpret that as being the new invisible hand, 0:11:30.040 --> 0:11:34.120 which we call artificial intelligence. Invoking an external force like 0:11:34.200 --> 0:11:38.479 the invisible hand, or an algorithm that automatically reads resumes 0:11:38.640 --> 0:11:43.560 or makes parole recommendations obsculls real human decisions. We have 0:11:43.640 --> 0:11:46.760 to remember that our creations reflect us. If you use 0:11:46.840 --> 0:11:50.040 that to abdicate your responsibility. If you use it just 0:11:50.080 --> 0:11:52.400 to cower and fear, then you're not being a good 0:11:52.400 --> 0:11:56.040 computer scientist. That is not the responsible way to do things. 0:11:56.080 --> 0:11:59.199 Just as if an economist says, well, the invisible hand 0:11:59.240 --> 0:12:03.040 says all these people should starve, that's not a responsible economist. 0:12:03.080 --> 0:12:07.040 The responsible economist fixes the problem in a sense. I 0:12:07.040 --> 0:12:11.520 think it's very hard to be effective if you believe 0:12:11.640 --> 0:12:14.719 in some kind of magical agency in your own inventions. 0:12:14.760 --> 0:12:16.720 I think you make yourself into an idiot. And and 0:12:16.800 --> 0:12:20.360 so I'm really concerned that not only economists but computer 0:12:20.440 --> 0:12:24.200 scientists make that error all the time. It's almost like 0:12:24.240 --> 0:12:27.480 a new form of mythology. I've been calling it alchemy lately. 0:12:27.520 --> 0:12:30.920 But yeah, sure, it's certainly easier to say, oh, we 0:12:30.960 --> 0:12:34.160 should respect this amazing autonomous living thing that has arisen 0:12:34.160 --> 0:12:36.719 in our own inventions. It's much easier to say that 0:12:36.720 --> 0:12:39.920 when it's benefiting you and you're getting very rich. Jarn 0:12:40.000 --> 0:12:42.800 puts his finger on a central irony in our relationship 0:12:42.840 --> 0:12:46.760 with technology. When our creations benefit us, we're quick to 0:12:46.800 --> 0:12:51.440 forget who pays the price. People who translate between natural languages, 0:12:51.480 --> 0:12:54.920 such as between English and Spanish, have seen their career 0:12:54.960 --> 0:12:59.599 prospects on the whole decrease tenfold since the arrival of 0:12:59.640 --> 0:13:02.679 Autumn Addic Translation, which is offered for free but companies 0:13:02.679 --> 0:13:06.960 like Google and Microsoft. Now, the thing is, you might say, well, 0:13:07.080 --> 0:13:09.560 this is very sad, but it always happens. People are made, 0:13:09.679 --> 0:13:14.160 people's jobs become obsolete when new technologies come along. The 0:13:14.200 --> 0:13:17.560 buggy whip goes away and the motor car comes. All right, 0:13:17.960 --> 0:13:20.600 but the problem is that every single day, those of 0:13:20.679 --> 0:13:24.480 us who help run these free services have to scrape 0:13:24.559 --> 0:13:27.439 or steal tens of millions of example phrases from all 0:13:27.440 --> 0:13:29.960 over the world from people who don't know it's being 0:13:29.960 --> 0:13:32.600 done to them. And the reason why every single day 0:13:32.640 --> 0:13:35.439 there's new pop culture and slang in public events and 0:13:35.920 --> 0:13:38.240 memes and on and on and so you need to 0:13:38.320 --> 0:13:41.720 constantly get new phrase examples to feed into the translation engines. 0:13:42.160 --> 0:13:44.160 So it's a weird thing. We're telling the people you 0:13:44.200 --> 0:13:46.360 don't get a job anymore because you're not needed. Oh, 0:13:46.400 --> 0:13:48.400 by the way, you're needed. We need to steal with you. 0:13:48.440 --> 0:13:50.120 Oh but by the way, we won't even tell you. 0:13:50.360 --> 0:13:52.360 And it's all based on this lie that we don't 0:13:52.360 --> 0:13:55.120 need people. Um, and that lie is based on this 0:13:55.200 --> 0:13:57.600 need to pretend the AI is this free standing thing, 0:13:57.760 --> 0:14:00.680 whereas we could instead think of it it's just the 0:14:00.720 --> 0:14:03.640 way that's African channel value between people in a new 0:14:03.679 --> 0:14:07.520 and better way. The free availability of real time translation 0:14:07.960 --> 0:14:12.320 opens up a world of possibilities for travelers, for language learners, 0:14:12.360 --> 0:14:17.240 for long distance lovers. But these technologies have invisible costs, 0:14:17.240 --> 0:14:19.600 like the translators losing their jobs to a tool that 0:14:19.720 --> 0:14:22.880 was trained on their work. And this kind of unpaid 0:14:22.960 --> 0:14:27.320 labor is actually something all of us participating every day 0:14:27.360 --> 0:14:31.280 without even realizing. Here's Lisa again. The way that these 0:14:31.440 --> 0:14:35.440 voice activated assistance are being trained is through huge amounts 0:14:35.520 --> 0:14:39.520 of data. A bit of an unknown secret by many 0:14:39.560 --> 0:14:42.760 people who have an Alexa is that every single time 0:14:42.920 --> 0:14:46.480 you are talking to that device, it's being recorded and 0:14:46.560 --> 0:14:49.560 being stored and going back to the cloud to train 0:14:49.880 --> 0:14:53.800 all of the other echoes around the world. So the 0:14:54.000 --> 0:14:57.960 users of echo devices are providing free labor on behalf 0:14:57.960 --> 0:15:02.480 of these massive organizations in order to train the system. 0:15:02.520 --> 0:15:05.200 It's not the first thing we think that after purchasing 0:15:05.240 --> 0:15:08.520 an Alexa and using it to buy stuff online, every 0:15:08.520 --> 0:15:11.520 time we interact with it, we're also helping Amazon improve 0:15:11.720 --> 0:15:14.840 and make more money, but framing our data in terms 0:15:14.840 --> 0:15:17.680 of labor helps us think about technology in new ways, 0:15:18.360 --> 0:15:22.360 and our Alexa use reminds Jarren of another science fiction movie. 0:15:22.800 --> 0:15:24.520 The reason. One that's really gotten to me is that 0:15:24.800 --> 0:15:29.160 probably the most famous cautionary tale about computers was in 0:15:29.280 --> 0:15:32.200 two thousand one Stanley Kubrick and Arthur C. Clark's movie 0:15:34.040 --> 0:15:36.600 And There's this computer called how that's this round thing 0:15:36.680 --> 0:15:38.280 that sits on the wall and just looks at you 0:15:38.400 --> 0:15:41.640 and talks to you, I'm sorry, I'm afraid I can't 0:15:41.680 --> 0:15:44.040 do that, and it ends up going berserk and killing 0:15:44.120 --> 0:15:46.080 people and just and they have to the program it. 0:15:46.280 --> 0:15:48.480 And the hot new gadge for the last few years 0:15:48.520 --> 0:15:50.240 has been this round thing that sits there and looks 0:15:50.280 --> 0:15:54.200 at you and you talk to I'm afraid I can't 0:15:53.520 --> 0:15:57.120 do these smart speakers and so on, and it's like, 0:15:57.280 --> 0:15:59.600 no matter how many cautions we put forward, people just 0:15:59.680 --> 0:16:02.880 follow right into it. It's it's astonishing to me. Of course, 0:16:02.920 --> 0:16:06.200 Alectra is powered by artificial intelligence. It takes a machine 0:16:06.280 --> 0:16:09.320 learning to understand what you say to it. But maybe 0:16:09.320 --> 0:16:12.000 the bigger breakthrough has been our decision to let listening 0:16:12.000 --> 0:16:15.200 devices into our homes. Yeah, I think that's true. I 0:16:15.240 --> 0:16:17.680 didn't grow up talking to something in my house. It's 0:16:17.680 --> 0:16:21.160 interesting when you read those articles like deep inside North Korea. 0:16:21.720 --> 0:16:24.760 You know, the thing that journalists always writes, there's a 0:16:24.800 --> 0:16:28.080 speaker in every house which projects the chairman's voice into 0:16:28.080 --> 0:16:31.400 the homes. As always the shocking detail. And of course 0:16:31.440 --> 0:16:33.760 now we will have Alexas in our houses, you know. 0:16:33.800 --> 0:16:36.520 I think it's interesting that Amazon was being delivered to 0:16:36.600 --> 0:16:38.120 us the boxes on our door step. That was the 0:16:38.120 --> 0:16:41.520 farthest they were going to get right, And now with 0:16:41.720 --> 0:16:44.600 Echo and Dot, these are devices that are inside of 0:16:44.600 --> 0:16:47.320 our homes, that are on our countertops. We're at this 0:16:47.400 --> 0:16:50.920 place where Alexa is now a part of the family. 0:16:51.280 --> 0:16:53.880 And now we have this first generation of children growing 0:16:53.920 --> 0:16:57.400 up with Alexas and other smart devices at home, interacting 0:16:57.440 --> 0:16:59.800 with them, seeing their parents talk to them all the time, 0:17:00.360 --> 0:17:03.800 and they're already used to this responsive technology. The giving 0:17:03.840 --> 0:17:07.399 away our data piece is disconcerting, but there's another piece 0:17:07.440 --> 0:17:10.639 of our shifting relationship with technology, which is why I 0:17:10.720 --> 0:17:13.320 dragged Julian into New Jersey to one of the smartest 0:17:13.359 --> 0:17:16.399 homes I know. And no I'm not talking about i Q. 0:17:17.520 --> 0:17:19.800 Let me ask you a question. What does your little 0:17:19.840 --> 0:17:24.000 brother look like a little guy? So what would you say? 0:17:24.040 --> 0:17:32.119 What does Alexe look like? But where does she live in? 0:17:33.560 --> 0:17:37.080 They really so every time you talk to Alexei, you're 0:17:37.080 --> 0:17:42.800 talking to space, despite having an Alexa who lives in space. 0:17:43.280 --> 0:17:45.240 My friend and her husband live in the suburbs with 0:17:45.280 --> 0:17:47.959 their sons, who are almost two and five. When they 0:17:48.000 --> 0:17:50.439 recently moved to New Jersey to fit their expanding family, 0:17:50.560 --> 0:17:53.440 they did not scrimp on smart home devices. Yeah, we've 0:17:53.440 --> 0:17:56.200 got two little kids. We both work. It doesn't bother 0:17:56.280 --> 0:17:58.680 me that they know our habits make it easier for us, 0:17:58.720 --> 0:18:01.760 like to get stuff done. Like I'm cool with Amazon 0:18:01.880 --> 0:18:03.840 just sending me diapers because it knows what I need 0:18:03.840 --> 0:18:07.200 diapers for parents like my friends, devices like the Amazon Alexa, 0:18:07.440 --> 0:18:11.440 Google Home and their marriad counterparts are genuinely helpful, which 0:18:11.480 --> 0:18:14.520 is probably why over a hundred and eighteen million American 0:18:14.560 --> 0:18:17.760 households have a smart speaker. That's half of the US. 0:18:18.440 --> 0:18:20.520 And when you've got something so involved in your home 0:18:20.560 --> 0:18:23.280 life that it helps with diapers and groceries, it's bound 0:18:23.280 --> 0:18:25.600 to affect some other areas as well. Your dad said, 0:18:25.600 --> 0:18:30.959 they're too There's there are three women in the house Google, 0:18:31.440 --> 0:18:37.320 who I don't know, who's the third woman in the house? 0:18:43.520 --> 0:18:48.320 And even bedtime is different. Hey, Google, tell me a story. Sure, 0:18:48.520 --> 0:18:52.080 here's one from Nickelodeon. It was a sunny day and 0:18:52.200 --> 0:18:55.560 Mr Porter was visiting farmer. He'll have I'll Google reading 0:18:55.600 --> 0:18:57.199 a story when he's in bed and we don't want 0:18:57.200 --> 0:19:00.919 to read any more books. You for your Google go 0:19:01.040 --> 0:19:05.679 off when you as if closed the door. Yeah. To 0:19:05.760 --> 0:19:09.760 be clear, Google Home has not replaced real life bedtime stories, 0:19:10.400 --> 0:19:12.480 but it has enabled my friend's kid to get more 0:19:12.480 --> 0:19:15.600 out of bedtime. He can keep asking for stories long 0:19:15.640 --> 0:19:18.639 after his parents need to stop reading. But focusing on 0:19:18.680 --> 0:19:22.080 the privacy component of these devices doesn't capture the full picture. 0:19:22.600 --> 0:19:24.879 Not only can the rhythms of family life change in 0:19:24.920 --> 0:19:28.520 response to a digital assistant, but so can kids expectations. 0:19:28.960 --> 0:19:31.119 And that's true for all of us, even those of 0:19:31.200 --> 0:19:35.119 us already passed early childhood development. We should think about 0:19:35.119 --> 0:19:39.240 how we're affected long term by our expectation of seamless delivery. 0:19:40.680 --> 0:19:43.640 You've got a good rapport with those kids, Kara. Neither 0:19:43.680 --> 0:19:46.000 of us have our own kids, but our editor among 0:19:46.040 --> 0:19:48.600 us does. I was curious for his take on next 0:19:48.600 --> 0:19:51.320 to joining the family So I'd gone out to this 0:19:51.320 --> 0:19:54.360 wedding a couple of years ago in Seattle, and it 0:19:54.520 --> 0:19:58.040 was in this fancy hotel and there was an Alexa 0:19:58.080 --> 0:20:01.160 there in the room, and Azzie and I went out 0:20:01.280 --> 0:20:03.159 for dinner or something, and we left the kids with 0:20:03.200 --> 0:20:07.200 a sitter, and the kids were mesmerized because they never 0:20:07.320 --> 0:20:10.679 encountered one of these devices before, so they were watching 0:20:10.680 --> 0:20:13.440 the babysitter interact with it, calling up music and whatever else. 0:20:13.480 --> 0:20:16.280 But then she also ordered food, so they were just 0:20:16.320 --> 0:20:18.760 floored by this. Then the next week, we were back 0:20:18.760 --> 0:20:21.000 at home and I was watching my four year old 0:20:21.040 --> 0:20:23.680 just like stomping around the house, and she started barking 0:20:24.359 --> 0:20:31.480 Alexa pizza, and it was just so confusing. She immediately 0:20:31.520 --> 0:20:34.119 knew that there was this thing you could just bark 0:20:34.160 --> 0:20:37.520 at it and and get food, and she wanted results. 0:20:37.600 --> 0:20:41.199 But I have a conflicted feeling about all of this 0:20:41.400 --> 0:20:44.960 because I had grown up in the States, very middle class, 0:20:45.160 --> 0:20:47.080 and every couple of years we go to India and 0:20:47.160 --> 0:20:49.960 visit my relatives who were a little wealthier. And we 0:20:50.000 --> 0:20:51.960 went to a party once with one of my cousins 0:20:52.640 --> 0:20:55.680 and I saw this kid who was super wealthy, and 0:20:55.760 --> 0:20:58.680 he was yelling at his chauffeur. He was yelling at 0:20:58.720 --> 0:21:02.320 his mom, he was barking at the maid that they had, 0:21:02.440 --> 0:21:05.840 and it was just so gross. And when you see 0:21:06.000 --> 0:21:09.080 this sort of entitlement in front of you and expressed 0:21:09.080 --> 0:21:11.639 in this way, you don't want your kids growing up 0:21:11.680 --> 0:21:14.280 with that, right, and you want everyone to be treated 0:21:14.280 --> 0:21:16.320 as humans. And and so for my daughter to be 0:21:16.320 --> 0:21:20.440 stomping around barking what she wanted, you know, I don't 0:21:20.440 --> 0:21:24.040 want that to be her way of speech. This kind 0:21:24.040 --> 0:21:27.040 of reminds me of the movie Invasion of the Body Snatchers, 0:21:27.680 --> 0:21:29.760 except the Monash version, which is that you know, he 0:21:29.800 --> 0:21:32.280 went out to dinner with his wife. He comes back, 0:21:32.520 --> 0:21:35.720 his daughter is like Alexa pizza, and he's like, what 0:21:35.880 --> 0:21:39.320 has Alexa done with my chid? You know, she was 0:21:39.359 --> 0:21:42.000 like the sweet little girl, and now she's like a 0:21:42.119 --> 0:21:45.280 child that's aware of an Alexa based on one encounter. Right. 0:21:45.880 --> 0:21:47.919 I feel like if an alien came down from space 0:21:48.240 --> 0:21:50.160 and came into my apartment and saw me cooking dinner 0:21:50.160 --> 0:21:53.760 and then saw me go, Alexa turn on Paul Simon radio, 0:21:54.119 --> 0:21:56.159 the alien would be like, is she yelling at a 0:21:56.200 --> 0:21:58.920 woman who's going to go turn on the music? Um? 0:21:58.960 --> 0:22:02.040 I think right, now we are all, you know, Mangesh's 0:22:02.119 --> 0:22:06.639 daughter included, participating in this life altering moment of self 0:22:06.680 --> 0:22:10.240 delusion where we're sort of collectively accepting smart devices in 0:22:10.280 --> 0:22:13.320 our homes. We're not just accepting them, we're treating them 0:22:13.359 --> 0:22:15.880 as human right, And I think this is the slippery 0:22:15.960 --> 0:22:19.320 ish slope of using voice assistance. More and more we're 0:22:19.359 --> 0:22:22.920 speaking to voice activated devices as though there are family members, 0:22:23.520 --> 0:22:25.920 and there have got to be some long term implications, 0:22:25.960 --> 0:22:29.600 both emotional and psychological, of blurring the line. I agree, 0:22:29.920 --> 0:22:31.800 and we're going to hear more from Jared about that 0:22:31.840 --> 0:22:35.040 after the break, But we'll also speak to Poppy Crumb again, 0:22:35.160 --> 0:22:38.280 who believes that we've barely scratched the surface of how 0:22:38.320 --> 0:22:48.520 these devices can change how we live for the better children. 0:22:48.520 --> 0:22:51.879 Interacting with Alexa and Siri and other digital assistance is 0:22:51.920 --> 0:22:56.240 particularly striking because of the developmental implications, But interacting with 0:22:56.359 --> 0:23:00.840 machines who understand and respond to us can his questions 0:23:00.880 --> 0:23:05.199 about who and what gets to be treated as a person. Here, Jarren, 0:23:05.960 --> 0:23:10.320 I think the problem isn't the math or the computer 0:23:10.359 --> 0:23:13.600 science algorithms. I think the problem is our framework for 0:23:13.680 --> 0:23:17.040 thinking about them tends to be machine centric. Instead of 0:23:17.119 --> 0:23:19.320 human centric, and it tends to create dangerous for the 0:23:19.359 --> 0:23:21.439 whole and and to create a lot of confusion, And 0:23:21.480 --> 0:23:23.439