Toyota Built a $10 Billion Surveillance City

0:00 Toyota built a $10 billion private utopia tech city.

0:04 Last week, RS Technica got an inside tour of Toyota's woven city.

0:09 The 10 billion dollar private city of the future sounds very Disney.

0:13 The company has been in Japan has built in Japan.

0:17 Yeah, there we go.

0:18 That makes a lot more sense.

0:19 Uh and the picture they painted was a mix uh was mixed at best.

0:24 Toyota chairman Akio Toyota uh first announced woven city

0:28 at CES 2020 as part of a pivot from car maker

0:31 to mobility company and the first 100 residents handpicked alpha

0:37 testers that Toyota calls Weavers moved in 6 months ago.

0:42 Despite the city branding,

0:44 only about 10% of the planned 175 acre site is actually built so far,

0:49 although these things take time.

0:50 residents act as alpha and beta testers for a rotating slate of Toyota tech.

0:57 Interesting.

0:58 Toyota's stated rationale for the heavy surveillance is

1:02 yeah, this is where things get a little dystopian.

1:05 The with the woven city telling ours uh that onboard car

1:11 sensors can't reliably spot a kid darting out from behind a truck.

1:16 So, street level cameras everywhere is part of how

1:20 Toyota plans to hit its zero accidents goal.

1:24 In practice, the camera density is striking with ours counting eight

1:29 at a single intersection and a halfozen even in small coffee shops,

1:34 all feeding into Toyota's AI vision engine,

1:37 which can track people across cameras based on their clothing.

1:41 Toyota built a cons a consent system called data

1:44 fabric that lets residents opt in or out per service.

1:48 And 98% of weavers, I'm paraphrasing, because they were handpicked,

1:55 have agreed to camera equipped robots in their apartments.

2:00 I guess we all have robot vacuums.

2:02 Not all of us, but a lot of people do.

2:04 Mine's been broken for a year, but it is what it is.

2:06 Uh the company's longerterm plan is to refine the AI vision

2:09 engine inside Woven City and then sell it to actual municipalities.

2:16 Toyota also confirmed that chairman Ao Toyota has been

2:20 uploaded to the city's cloud as AO Toyota AI,

2:28 a chatbot trained on a decade of his speeches and writings.

2:32 Really?

2:38 Wow.

2:38 I think this has got to be

2:43 Wow.

2:44 Wow.

2:44 We have no idea what to spend our copious amounts of money on.

2:51 Wow.

2:52 So, let's do this.

2:53 And this is coming from someone who just attended an event

2:58 that was held in an abandoned power plant because a billionaire saw

3:09 that it was like kind of an eyes sore for his neighbors

3:12 and decided to buy it to turn it into a park.

3:19 Wow.

3:20 Very nice.

3:22 Yeah.

3:23 This is like this is actually peak.

3:27 My my my steak is too juicy.

3:30 My my family is too perfect.

3:34 My money my money bushels are overflowing.

3:38 I simply have no idea what to do with it.

3:40 This is like, you know, reminds me a lot of what what was the Disney thing?

3:47 What was the Disney thing?

3:48 The like city Disney city thing.

3:50 Oh, the town of tomorrow, right?

3:53 Yeah.

3:53 Yeah.

3:53 Yeah.

3:54 Disney.

3:55 This feels like like modern that Epcot.

3:59 That's the one.

4:00 Epcot.

4:00 Yeah.

4:01 Yeah.

4:01 Yeah.

4:05 Yeah.

4:04 Did Epcot have like insane surveillance?

4:07 I don't know if I don't know if our corporate overlords had imagined that yet

4:11 cuz this to me feels like I know what you're saying.

4:14 We don't know what to spend our money on.

4:16 I don't think that's true at all.

4:17 To be honest, I I don't get that read from this in the slightest.

4:20 I really, no, not even a little bit.

4:21 I get flock from this.

4:23 I get uh Toyota attempting to make technology to sell the governments,

4:28 which is literally what they're saying.

4:29 They're planning on selling this to municipalities.

4:32 Uh they want to be flock.

4:34 They want to have AI enabled cameras all over the place,

4:38 tracking everyone, measuring everything,

4:40 and then selling it to governments and probably

4:42 also companies for advertisement and everything else.

4:45 I this Yeah, I I don't get the same read as you on this.

4:49 So, so basically it sounds like because we're

4:53 neither of us is what they're saying it's for.

4:56 They're saying it's for, you know, connected cities for like, you know, reducing

5:00 No, they're saying it's for what I said.

5:03 Well, no.

5:03 I mean, they didn't say anything about crime.

5:05 They said, you know, like car accidents basically.

5:07 Yeah.

5:07 But they said they're going to sell it to municipalities.

5:10 No, no, I I No, I got that part.

5:12 But like flock is is uh is is for surveillance and and like uh suspect tracking.

5:20 Sell it to municipalities.

5:21 What are you going to be used for?

5:22 No, no, I know.

5:23 I know.

5:23 But the part is that I said they didn't say that though.

5:25 They did.

5:26 So they haven't said the quiet part out loud.

5:28 What they're saying is we're a mobility company

5:31 and this is about mobility and making it

5:34 like so that every car it has a brain on board it that is being fed

5:39 data from these AI enabled cameras so

5:43 that it would never like accidentally hit a kid

5:45 who's like darting out from between two cars

5:47 or you know chasing a ball or whatever.

5:49 Um, so your theory then is that whether Toyota intends it or not,

5:55 this will be this will be sold as a as a as just a mass surveillance tool.

6:01 Yeah.

6:01 I mean, cuz it the the whole cars angle doesn't make sense

6:04 when there's cameras inside of coffee shops and inside of your home.

6:08 So, I I read the RS article and um they

6:13 go into a little bit more depth about how, you know,

6:17 Toyota's culture and Japanese culture and and like privacy are

6:22 are sort of really central to this and what their plan

6:25 is for anonymizing things like that's that's part of why

6:29 the individual tracking is based on the clothing that you're wearing.

6:32 It does not use facial identification for instance.

6:37 However, I think that anyone as technical as anybody at Toyota working

6:43 on this project or anyone from ours Technica or anyone in our audience

6:47 is going to know that there's it's like just because you're not using

6:51 facial tracking doesn't mean that you can't pin this down to an individual,

6:54 especially if you've got a large enough set of data

6:58 monitoring like a fixed number of local residents, right?

7:01 Like whether we're talking about the outfits that they

7:04 normally wear or whether we're talking about gate analysis,

7:07 looking at how they walk, right?

7:09 Like it's um there's a lot of ways like

7:11 an actually pretty wild amount of ways to identify people.

7:14 Tell me this.

7:17 Okay, this is a hypothetical to be very very clear.

7:20 Tell me this.

7:22 If Toyota could solve privacy for this, if their pitch that this is

7:29 a mobility solution that does not store in any accessible way,

7:35 this is a huge if, I know,

7:37 does not store in any accessible way personally identifiable information.

7:42 And it is purely a machine vision system that sends feedback to to cars

7:49 and bikes and and these like smart equipped vehicles to prevent accidents.

7:54 Is this something that you could see having any kind of future?

7:58 I'm I'm going to be I need you to understand

8:01 I know I'm going to be super annoying for a second.

8:06 One thing is you said they if even if they don't store

8:08 the data and stuff the problem is there's other types of attacks.

8:12 So like it will be a privacy nightmare because everything

8:17 if they could solve it if if I know this is an if.

8:20 But the problem is that if is impossible.

8:23 I know but try to try try try to try to play

8:27 if it is the world's perfect uh given by some deity not of the earth

8:35 um data that's what we're assuming

8:37 that is completely unhackable that is completely literally

8:41 impossible then yeah I mean I don't I don't know if I really care um well for me

8:46 if it if it can never be if it is completely

8:49 uncorruptible and is only ever used for good and is completely unhackable,

8:54 then it's like, okay,

8:57 so I'm actually going to argue against my point now.

9:01 Okay.

9:00 And I'm going to say it has no future because then nobody would want it.

9:05 It would be too expensive to implement purely to save a handful of lives.

9:09 I mean, think about how many intersections you drive

9:11 through in a typical city that obviously need better signage,

9:15 but because that would cost a few hundred, it's not going to get implemented.

9:19 So this as like Toyota imagining this mobility future

9:24 where everything every object in a city is tracked.

9:28 I think even if they could solve it but this is why I wanted to have

9:32 the conversation right because you know it has to kind of boil down to like

9:36 if their pitch is right is there a future and I still think the answer is no.

9:41 Yeah.

9:41 Yeah.

9:42 because

9:42 I'd love to be wrong

9:43 and and you know people people can say what they want but ultimately when

9:48 it comes down to municipal budgets um what are people going to vote for?

9:54 I mean uh looking at like Langley right now

9:57 for example we're we're we're doing some weird stuff.

10:03 Hey Luke.

10:04 Okay, hold on a second though.

10:05 Hold on.

10:06 Yeah, hold on.

10:06 Let's um let's back up a little bit here because

10:08 I actually might be about to argue against myself again.

10:12 What if we what if we didn't do this with cameras?

10:14 Like what if we could do this with a point cloud device?

10:17 Like what if we could do this with LAR?

10:19 You'd still be able to Oh, man.

10:22 No.

10:22 Because you'd still be able to see a blob coming out of this square blob

10:29 at this time every day and going

10:30 to the other square blob at that time every day.

10:32 Yeah, never mind.

10:33 I talked myself out of it again.

10:36 Because like to to me like the you know

10:38 the cameras and the data analysis and no matter how much

10:41 comput and no matter how cheap cameras they're going

10:43 to get they're always they're always going to have some cost.

10:45 You also just can't get around it.

10:46 Like if if surveillance is being done,

10:50 it it can be hacked and it you know on a technical level it can be hacked.

10:57 On a social level it will almost certainly be corrupted.

11:01 There's like there's every example of, you know,

11:03 police officers using tracking equipment to track their girlfriends

11:07 to see if they're cheating or or or yeah,

11:10 following ex-girlfriends and ex-wives and whatnot.

11:12 There's there's every example of all these different

11:14 things where if it if it's a social system,

11:16 it can and almost certainly will be corrupted.

11:18 And if it's a technical system, it can be hacked.

11:21 And those things just like seem to effectively be laws of the universe.

11:24 So, it's tough when you have a system like

11:26 this because you just know both of those things are true.

11:29 I'm going to throw another wrench into the gears.

11:32 Yet another wrench because I just I

11:34 think this is just a really interesting conversation.

11:38 It's not interesting enough for me to have spent my 10 billion to, you know,

11:43 prove that it's not going to work or whatever.

11:46 I'm But I'm glad Toyota did it because Yeah.

11:51 Yeah.

11:51 Sure.

11:52 Um, I I I I have a 10 billion offer, $10 billion offer in to buy the city.

11:58 All they need is a $10 billion loan,

12:00 but but TD TD seems like their game for this sort of thing.

12:04 So, uh, they totally totally do it.

12:06 Um, so yeah, let me let me throw this out and go, okay, but hear me out.

12:12 As our methods for tracking these things get

12:17 more and more ubiquitous and more and more

12:20 precise and more and more absolutely everybody has

12:24 them on their person at any given moment,

12:30 how different would the point cloud system be?

12:35 I guess you could make a strong you can make

12:37 a strong argument that the camera system is is different, but like

12:43 it how different would it be?

12:45 Hit me up.

12:46 I mean, that's already a problem, right?

12:48 I I think we talked on a fairly recent WAN show about

12:51 a crime scene where Google got what is the right term for it?

12:55 Subpoenaed or something.

12:56 Uh for information on what devices were in the area

12:59 at the time in this like big window.

13:02 Um, so they were poking around in the phones of people and we know from

13:07 there was no reason to suspect them.

13:08 We know from the Snowden era, I don't know if he's still doing public stuff, uh,

13:12 but we know from when he was like a a big

13:14 deal in the news um that when when he was roaming around,

13:17 you know, he's throwing phones in in uh freezers and stuff like

13:22 that to try to Faraday cage them um because they were already a problem.

13:26 So, it's like it's it yeah, I mean that's definitely part of it,

13:30 but it's I don't know adding more fuel

13:34 to the fire because the phones aren't perfect.

13:35 You could decide to leave it at home.

13:38 Um I remember when we talked about that story of the Yeah,

13:40 someone said geofencing warrant.

13:42 When we talked about the story of the geoencing warrant,

13:44 it was kind of like astonishing to me that you would

13:46 bring your cell phone to a crime scene because it's like,

13:48 yeah, I mean, you can obviously be tracked.

13:51 What are you doing?

13:53 Um, well, that's the thing with criminals, Luke.

13:58 Yeah.

13:58 Yeah.

13:59 Um, so yeah, it's tough.

14:02 It's tough.

14:03 It's also ultimately our opinion on it isn't really going to matter, I think.

14:09 Uh, like every single person I've ever

14:12 heard of who has talked about flock in any way has been like, "Wow,

14:16 this is obviously bad and evil and a terrible system." Um,

14:20 and they're all over the place.

14:22 So, I don't know.