WEBVTT - The Scary Implications of Facial Recognition Technology

0:00:00.120 --> 0:00:03.960
<v Speaker 1>This is Bloomberg Business Week with Carol Messer and Tim

0:00:04.000 --> 0:00:07.560
<v Speaker 1>Steneveek on Bloomberg Radio.

0:00:09.080 --> 0:00:17.160
<v Speaker 2>You love Love Hey like it or not? Everybody facial

0:00:17.200 --> 0:00:18.239
<v Speaker 2>recognition technology.

0:00:18.280 --> 0:00:18.880
<v Speaker 3>It is everywhere.

0:00:18.880 --> 0:00:21.400
<v Speaker 2>You may not realize it, but it is everywhere. It uses,

0:00:22.520 --> 0:00:24.840
<v Speaker 2>you know, in terms of the applications vary from the

0:00:24.920 --> 0:00:27.720
<v Speaker 2>mundane to well, you know, we use it right to

0:00:27.720 --> 0:00:29.920
<v Speaker 2>open up our phone, yeah, open up other stuff. But

0:00:30.000 --> 0:00:32.880
<v Speaker 2>it also the uses can be downright controversial.

0:00:32.960 --> 0:00:35.199
<v Speaker 1>I mean just think about Madison Square Garden. Yes, you've

0:00:35.240 --> 0:00:38.479
<v Speaker 1>heard about this. It reportedly uses facial recognition technology to

0:00:38.520 --> 0:00:41.120
<v Speaker 1>weed out lawyers at firms that are suing the company

0:00:41.479 --> 0:00:45.159
<v Speaker 1>and bevets them from going to concerts and shows at

0:00:45.159 --> 0:00:45.640
<v Speaker 1>the arena.

0:00:45.720 --> 0:00:46.680
<v Speaker 2>Welcome to the new reality.

0:00:46.760 --> 0:00:48.640
<v Speaker 1>Yeah, that's what it is. What can go wrong though,

0:00:48.640 --> 0:00:51.519
<v Speaker 1>when law enforcement things or oppressed the governments get their

0:00:51.520 --> 0:00:54.200
<v Speaker 1>hands on this technology. What happens when the technology gets

0:00:54.200 --> 0:00:55.760
<v Speaker 1>so good that a company claims it can take a

0:00:55.800 --> 0:00:58.600
<v Speaker 1>single photo of someone and then use that photo to

0:00:58.600 --> 0:01:00.880
<v Speaker 1>find every detail of their life on the Internet.

0:01:00.960 --> 0:01:02.880
<v Speaker 2>It's the premise of a new national best selling book

0:01:02.880 --> 0:01:05.640
<v Speaker 2>by New York Times technology reporter kashmiir Hill. The book

0:01:05.680 --> 0:01:08.480
<v Speaker 2>is called your face belongs to us, A secretive startups

0:01:08.520 --> 0:01:10.600
<v Speaker 2>quest to end privacy as we know it, and is

0:01:10.840 --> 0:01:12.800
<v Speaker 2>no doubt about it. Tim, A very timely conversation.

0:01:12.880 --> 0:01:13.039
<v Speaker 4>Yeah.

0:01:13.040 --> 0:01:15.120
<v Speaker 1>The book has been longlisted for the Ft and Schroeder's

0:01:15.160 --> 0:01:17.240
<v Speaker 1>Business Book of the Year awards. Kashmir right now joins

0:01:17.280 --> 0:01:19.440
<v Speaker 1>us on Zoom from New York. Kashmir has been covering

0:01:19.440 --> 0:01:21.880
<v Speaker 1>the intersection of privacy and tech for around fifteen years

0:01:21.920 --> 0:01:24.400
<v Speaker 1>before people were even thinking about privacy online. So very

0:01:24.400 --> 0:01:27.600
<v Speaker 1>pleased to have her with us this afternoon. Kashmir, good

0:01:27.600 --> 0:01:30.480
<v Speaker 1>to talk to you. The book centers around Clearview AI.

0:01:30.600 --> 0:01:32.760
<v Speaker 1>It's the secretive startup which you've been writing about and

0:01:32.800 --> 0:01:36.040
<v Speaker 1>following for years. It's got some pretty incredible technology. Tell

0:01:36.120 --> 0:01:36.679
<v Speaker 1>us what they do?

0:01:38.200 --> 0:01:40.360
<v Speaker 3>Yeah, Hi Tim, Hi Carol, good to be here.

0:01:40.640 --> 0:01:44.720
<v Speaker 4>So what CLEARVIEWI did was it went out and scraped

0:01:44.880 --> 0:01:48.680
<v Speaker 4>billions of photos from the Internet from social media sites

0:01:48.760 --> 0:01:53.960
<v Speaker 4>like Facebook, Instagram, LinkedIn, Venmo. It did that without anyone's consent.

0:01:54.120 --> 0:01:58.280
<v Speaker 4>Now has a database of thirty billion faces and has

0:01:58.320 --> 0:02:01.120
<v Speaker 4>this facial recognition app where you can take a photo

0:02:01.160 --> 0:02:03.960
<v Speaker 4>of someone and it will bring up all these other

0:02:04.040 --> 0:02:07.520
<v Speaker 4>places on the Internet where it's seen their face. Along

0:02:07.560 --> 0:02:09.360
<v Speaker 4>with a link to where it appears, so you can

0:02:09.400 --> 0:02:12.320
<v Speaker 4>find out someone's name, you know, where they live, who

0:02:12.400 --> 0:02:14.880
<v Speaker 4>they know, maybe photos. They didn't even know we're on

0:02:14.919 --> 0:02:18.600
<v Speaker 4>the internet. And they've been selling that app to police

0:02:18.639 --> 0:02:20.280
<v Speaker 4>now for a few years.

0:02:22.200 --> 0:02:24.600
<v Speaker 1>Okay, So that's a really important point. It's not like

0:02:24.639 --> 0:02:26.640
<v Speaker 1>you can go download this app from the app store

0:02:26.720 --> 0:02:29.200
<v Speaker 1>and just start, you know, looking at photos of your

0:02:29.200 --> 0:02:32.080
<v Speaker 1>colleagues and your friends on there. You've been able to

0:02:32.320 --> 0:02:36.120
<v Speaker 1>see it in action though. Kashmir from the company tell

0:02:36.200 --> 0:02:37.960
<v Speaker 1>us what it's like to kind of experience that.

0:02:40.280 --> 0:02:42.760
<v Speaker 4>Yeah, when I first started looking into Clearview AI, they

0:02:42.800 --> 0:02:45.720
<v Speaker 4>were not happy about it, and they actually put an

0:02:45.720 --> 0:02:48.120
<v Speaker 4>alert on my face so that they would know when

0:02:48.200 --> 0:02:51.280
<v Speaker 4>police officers were I was talking to we're searching my face,

0:02:51.560 --> 0:02:55.320
<v Speaker 4>and they blocked my face from having results, which was

0:02:55.360 --> 0:03:00.200
<v Speaker 4>which was a bit chilling. But eventually the come and

0:03:00.200 --> 0:03:03.680
<v Speaker 4>he started talking to me, and its founder, Wanton Tat,

0:03:04.480 --> 0:03:07.520
<v Speaker 4>has run searches on me several times over these years

0:03:07.560 --> 0:03:09.600
<v Speaker 4>now that I've been talking to him for my reporting

0:03:09.600 --> 0:03:11.840
<v Speaker 4>for The Times and for the book, and I just

0:03:11.919 --> 0:03:14.560
<v Speaker 4>remember the first time he did this search of my

0:03:14.680 --> 0:03:17.960
<v Speaker 4>face and you know, it brought up photos I knew about,

0:03:18.120 --> 0:03:21.240
<v Speaker 4>you know, you know, headshots that I have put out

0:03:21.240 --> 0:03:23.880
<v Speaker 4>there in the world. But it was also photos I

0:03:23.880 --> 0:03:25.720
<v Speaker 4>had never seen before, photos of me.

0:03:26.080 --> 0:03:28.560
<v Speaker 3>With a source that I was interviewing for a story.

0:03:28.600 --> 0:03:31.440
<v Speaker 4>We were out at a public event, a photo where

0:03:31.600 --> 0:03:34.639
<v Speaker 4>it was somebody standing in Washington, d C. With somebody

0:03:34.680 --> 0:03:37.160
<v Speaker 4>walking by in the background. And when I first looked

0:03:37.160 --> 0:03:38.680
<v Speaker 4>at the photo, I didn't think I was in it

0:03:38.880 --> 0:03:42.280
<v Speaker 4>until I recognized the jacket the person in the background

0:03:42.360 --> 0:03:44.600
<v Speaker 4>was wearing, and it was a very distinctive coat that

0:03:44.640 --> 0:03:47.480
<v Speaker 4>I had bought in Tokyo. And so, you know, when

0:03:47.520 --> 0:03:50.960
<v Speaker 4>this technology works, it can be incredible the photos that

0:03:50.960 --> 0:03:51.880
<v Speaker 4>it's able to pull up.

0:03:53.800 --> 0:03:56.119
<v Speaker 2>Cash I would say that I think we all realize.

0:03:56.200 --> 0:03:58.640
<v Speaker 2>I mean, I've got a security badge that I have

0:03:58.720 --> 0:04:01.400
<v Speaker 2>to check into my office, right, I know that as

0:04:01.400 --> 0:04:04.280
<v Speaker 2>I walk around the world out on streets, you know,

0:04:04.280 --> 0:04:06.240
<v Speaker 2>you go over in Europe, like their cameras are everywhere.

0:04:06.280 --> 0:04:08.240
<v Speaker 2>So I'm I have an awareness, and I think a

0:04:08.280 --> 0:04:09.960
<v Speaker 2>lot of us do that. So much of what we

0:04:10.040 --> 0:04:12.960
<v Speaker 2>do is being tracked digitally, right, we have these dossiers.

0:04:13.000 --> 0:04:16.560
<v Speaker 2>Everything's being tracked. It's what people do with that information.

0:04:16.640 --> 0:04:19.039
<v Speaker 2>I feel like we're at this era of trying to

0:04:19.120 --> 0:04:22.120
<v Speaker 2>understand where that information goes, how it's used to either

0:04:22.760 --> 0:04:25.839
<v Speaker 2>you know, help or you know, more worrisome hurt us.

0:04:26.200 --> 0:04:30.159
<v Speaker 2>So what is it about this company is worrisome about

0:04:30.200 --> 0:04:31.760
<v Speaker 2>kind of maybe the use of that data.

0:04:33.120 --> 0:04:35.279
<v Speaker 4>Yeah, I mean, I think what's worrisome about it is

0:04:35.320 --> 0:04:37.960
<v Speaker 4>this idea that we were all put in this database.

0:04:38.000 --> 0:04:40.440
<v Speaker 4>I mean, I'm sure the three of us are in here, uh,

0:04:40.760 --> 0:04:43.360
<v Speaker 4>and that we never got a choice in the matter.

0:04:44.279 --> 0:04:47.080
<v Speaker 4>That it was operating secretly, you know, until I came

0:04:47.120 --> 0:04:49.599
<v Speaker 4>along and discovered it. We didn't even know police had

0:04:49.640 --> 0:04:53.680
<v Speaker 4>a technology like this. You know, the algorithm that they

0:04:53.720 --> 0:04:56.560
<v Speaker 4>were using hadn't been tested. We'd actually didn't know how

0:04:56.600 --> 0:05:00.240
<v Speaker 4>accurate it was. And now that the technology is is

0:05:00.279 --> 0:05:03.760
<v Speaker 4>out there, I think it just raises concerns about our

0:05:03.800 --> 0:05:08.200
<v Speaker 4>ability to kind of move anonymously in a public space

0:05:08.400 --> 0:05:11.480
<v Speaker 4>and trust that people aren't going to know who we are.

0:05:11.880 --> 0:05:14.480
<v Speaker 3>Clear UAI kind of broke this taboo.

0:05:14.480 --> 0:05:16.920
<v Speaker 4>I write in the book it wasn't a technological breakthrough,

0:05:17.040 --> 0:05:20.080
<v Speaker 4>is actually an ethical one, and we are going to

0:05:20.120 --> 0:05:23.479
<v Speaker 4>see copycats and this will be available to all of us,

0:05:23.560 --> 0:05:25.719
<v Speaker 4>and it will mean that you can't have a private

0:05:25.760 --> 0:05:28.880
<v Speaker 4>conversation over dinner, and trust that the people around you

0:05:28.920 --> 0:05:32.599
<v Speaker 4>can't identify you, you can't buy something sensitive to pharmacy,

0:05:32.640 --> 0:05:35.159
<v Speaker 4>and trust that somebody sees you can't take your picture

0:05:35.160 --> 0:05:37.120
<v Speaker 4>and know who you are. It just could really change

0:05:37.400 --> 0:05:42.240
<v Speaker 4>our ability to just have you know, private going goings

0:05:42.279 --> 0:05:44.719
<v Speaker 4>on in the world where anybody can take our picture.

0:05:46.520 --> 0:05:48.479
<v Speaker 2>Just think of the story the Bloomberg Business Week's story

0:05:48.480 --> 0:05:51.880
<v Speaker 2>about location services with Google right, and police using those also.

0:05:52.040 --> 0:05:54.720
<v Speaker 1>It's just it's kind of the perfect segue to my

0:05:54.760 --> 0:05:57.880
<v Speaker 1>next question for Kashmir, which is which is why Kashmir

0:05:57.960 --> 0:06:01.600
<v Speaker 1>this is Clearview doing this and not Google or Facebook.

0:06:01.600 --> 0:06:03.960
<v Speaker 1>You spent a big part of the book talking about

0:06:04.000 --> 0:06:07.440
<v Speaker 1>that why it is Clearview even though technologically Google and

0:06:07.480 --> 0:06:09.200
<v Speaker 1>Facebook could do this. I mean, think about all the

0:06:09.200 --> 0:06:11.920
<v Speaker 1>pictures that Facebook has willingly from us, So Google, if

0:06:11.920 --> 0:06:13.880
<v Speaker 1>we use Google Photos, has from us.

0:06:15.680 --> 0:06:18.240
<v Speaker 4>Yeah. I mean when I first found out about Clearviewyi,

0:06:18.360 --> 0:06:21.360
<v Speaker 4>I just assumed that they were super geniuses, that they

0:06:21.360 --> 0:06:25.760
<v Speaker 4>had been able to accomplish this thing, that Facebook and

0:06:25.800 --> 0:06:28.120
<v Speaker 4>Google hadn't released, that governments hadn't built.

0:06:28.520 --> 0:06:30.560
<v Speaker 3>But I did find out in my reporting for.

0:06:30.520 --> 0:06:34.720
<v Speaker 4>The book that Google, Google and Facebook both developed this

0:06:34.839 --> 0:06:38.599
<v Speaker 4>kind of technology internally. Eric Schmidt as early as twenty

0:06:38.680 --> 0:06:42.320
<v Speaker 4>eleven s that it was the one technology that Google

0:06:42.360 --> 0:06:45.560
<v Speaker 4>had developed and decided not to release. I watched this

0:06:45.960 --> 0:06:49.800
<v Speaker 4>video of engineers at Facebook and twenty seventeen in this

0:06:49.839 --> 0:06:53.040
<v Speaker 4>little conference room and then Low Park with a smartphone

0:06:53.480 --> 0:06:56.239
<v Speaker 4>on the brim of a baseball cap held in place

0:06:56.279 --> 0:06:59.080
<v Speaker 4>by rubber bands, and when the person wearing it looked

0:06:59.080 --> 0:07:01.719
<v Speaker 4>around at people in the room, it actually.

0:07:01.320 --> 0:07:02.440
<v Speaker 3>Called out their names.

0:07:02.920 --> 0:07:05.520
<v Speaker 4>And this was really striking to me, because neither one

0:07:05.560 --> 0:07:08.640
<v Speaker 4>of these companies released the product, and they just felt

0:07:08.640 --> 0:07:11.040
<v Speaker 4>like it was too dangerous to put this out in

0:07:11.080 --> 0:07:13.280
<v Speaker 4>the world, that it could be used in a bad way.

0:07:13.720 --> 0:07:17.600
<v Speaker 4>And so yeah, it is very striking that Clearview was

0:07:17.680 --> 0:07:19.840
<v Speaker 4>the company to do it. And I think it's in

0:07:19.840 --> 0:07:23.640
<v Speaker 4>part because those companies Facebook, Google, they're just under more scrutiny.

0:07:23.680 --> 0:07:25.960
<v Speaker 4>They have more to lose, They've gotten in so much

0:07:25.960 --> 0:07:29.360
<v Speaker 4>trouble over the years for privacy violations, whereas clear Views,

0:07:29.440 --> 0:07:33.760
<v Speaker 4>this radical startup, you know, it has only only everything

0:07:33.800 --> 0:07:37.920
<v Speaker 4>to gain by kind of putting this this radical technology

0:07:37.920 --> 0:07:39.040
<v Speaker 4>out into the world.

0:07:40.360 --> 0:07:42.720
<v Speaker 2>I want to get into, like maybe the investigations in

0:07:42.800 --> 0:07:45.040
<v Speaker 2>terms of security forces around the world that maybe have

0:07:45.080 --> 0:07:46.920
<v Speaker 2>been looking into this company. Having said this, if you're

0:07:46.960 --> 0:07:49.400
<v Speaker 2>not already freaked out, we have about a minute and

0:07:49.400 --> 0:07:50.760
<v Speaker 2>a half and then we'll come back and talk some more.

0:07:50.800 --> 0:07:53.480
<v Speaker 2>But tell us about some of the individuals that have

0:07:53.600 --> 0:07:56.040
<v Speaker 2>somehow been connected with this company or behind this company.

0:07:57.360 --> 0:07:57.680
<v Speaker 3>Yeah.

0:07:57.720 --> 0:08:00.640
<v Speaker 4>So the main person I talked about before, Hwani ton Tat.

0:08:00.680 --> 0:08:03.760
<v Speaker 4>He is the technical co founder. Grew up in Australia,

0:08:03.920 --> 0:08:07.080
<v Speaker 4>loved technology nineteen years old, dropped out of college, moved

0:08:07.080 --> 0:08:09.320
<v Speaker 4>to San Francisco, was just kind of trying to make

0:08:09.360 --> 0:08:12.320
<v Speaker 4>it in the tech world, the tech gold rush in

0:08:12.840 --> 0:08:16.280
<v Speaker 4>kind of the early two thousands, and he ended up

0:08:16.320 --> 0:08:19.160
<v Speaker 4>moving to New York and really falling in with a

0:08:19.360 --> 0:08:22.280
<v Speaker 4>more conservative crowd, people that were kind of on the

0:08:22.280 --> 0:08:26.360
<v Speaker 4>alt right. He ended up partnering with Richard Schwartz, decades

0:08:26.400 --> 0:08:29.360
<v Speaker 4>older than him, had worked for Rudy Giuliani in New

0:08:29.400 --> 0:08:32.960
<v Speaker 4>York City when he was the mayor. Another person helped

0:08:32.960 --> 0:08:35.920
<v Speaker 4>them in the early days Charles Johnson, kind of well

0:08:35.960 --> 0:08:38.840
<v Speaker 4>known on the internet as a conservative provocateur, a little

0:08:38.840 --> 0:08:42.960
<v Speaker 4>bit of a troll of liberals, and they were working

0:08:42.960 --> 0:08:46.360
<v Speaker 4>on this together. Peter Thiel was their first investor, with

0:08:46.400 --> 0:08:49.400
<v Speaker 4>two hundred thousand dollars for the company that became CLEARVIEWAI.

0:08:49.880 --> 0:08:52.040
<v Speaker 3>It really kind of came out of those.

0:08:53.280 --> 0:08:58.320
<v Speaker 4>Those kind of online right currents around the time that

0:08:58.360 --> 0:08:59.640
<v Speaker 4>Trump was becoming president.

0:09:02.160 --> 0:09:04.400
<v Speaker 2>All right, we're going to continue with the conversation. I've

0:09:04.440 --> 0:09:07.400
<v Speaker 2>got to say, we're all just kind of hanging on

0:09:07.440 --> 0:09:09.440
<v Speaker 2>each word because I think it's you know, this is

0:09:09.480 --> 0:09:11.400
<v Speaker 2>the reality of our world. Right now, we have an

0:09:11.480 --> 0:09:13.760
<v Speaker 2>understanding of it, but you begin to wonder, Okay, how

0:09:13.760 --> 0:09:14.840
<v Speaker 2>does it get used exactly?

0:09:14.920 --> 0:09:16.320
<v Speaker 1>And when we come back, we're going to speak to

0:09:16.600 --> 0:09:19.960
<v Speaker 1>Kashmir Hill, the author of this new book it's called

0:09:20.000 --> 0:09:22.400
<v Speaker 1>Your Face Belongs to Us, A Secret of Startups, quest

0:09:22.440 --> 0:09:24.079
<v Speaker 1>and Privacy as we know it. We're going to dive

0:09:24.120 --> 0:09:29.320
<v Speaker 1>into the organizations that are using this technology, the law

0:09:29.400 --> 0:09:32.320
<v Speaker 1>enforcement organizations, then also talk about how this technology is

0:09:32.320 --> 0:09:34.520
<v Speaker 1>being used around the world by oppressive government.

0:09:34.559 --> 0:09:36.400
<v Speaker 2>Right. We all kind of google ourselves, let's be honest.

0:09:36.440 --> 0:09:38.360
<v Speaker 2>But like normal, things pop up, but if you start

0:09:38.400 --> 0:09:39.720
<v Speaker 2>to see all of a sudden, these pictures come up

0:09:39.720 --> 0:09:41.680
<v Speaker 2>and you're like, where did that come from? That would

0:09:41.679 --> 0:09:44.440
<v Speaker 2>freak me out. Heykash were one thing we're wondering, you know,

0:09:44.600 --> 0:09:46.280
<v Speaker 2>and you get into this in the book is just

0:09:47.559 --> 0:09:51.360
<v Speaker 2>you know, global forces, global regulators around the world. I feel,

0:09:51.440 --> 0:09:53.480
<v Speaker 2>especially when you go outside the United States, you go

0:09:53.480 --> 0:09:57.000
<v Speaker 2>over in Europe, they're really concerned privacy issues when it

0:09:57.040 --> 0:10:00.080
<v Speaker 2>comes to individuals. So what kind of scrutiny and what

0:10:00.160 --> 0:10:04.720
<v Speaker 2>kind of oversight has Clearview kind of garnered from some

0:10:04.760 --> 0:10:06.160
<v Speaker 2>of the regulators around the world.

0:10:07.360 --> 0:10:11.160
<v Speaker 4>Yeah, how private your faces really varies depending on where

0:10:11.200 --> 0:10:16.040
<v Speaker 4>you live. And so after I kind of expose the

0:10:16.080 --> 0:10:22.120
<v Speaker 4>existence of CLEARVIEWAI, privacy regulators in Europe, Canada, Australia all

0:10:22.160 --> 0:10:25.480
<v Speaker 4>announced investigations into the company, and they all came to

0:10:25.520 --> 0:10:30.160
<v Speaker 4>the same conclusion, which is that CLEARVIEWAI violated their privacy

0:10:30.240 --> 0:10:34.120
<v Speaker 4>laws by collecting all this data about their citizens, this

0:10:34.240 --> 0:10:38.680
<v Speaker 4>kind of sensitive biometric information without their consent, and they said, hey,

0:10:38.720 --> 0:10:41.920
<v Speaker 4>this is illegal. Some of the privacy regulators find them,

0:10:42.280 --> 0:10:45.280
<v Speaker 4>and Clearview had been offering you know, its product to

0:10:45.880 --> 0:10:49.480
<v Speaker 4>officers around the world, and this effectively kind of kicked

0:10:49.880 --> 0:10:51.440
<v Speaker 4>a Clearview out of those countries.

0:10:53.720 --> 0:10:54.800
<v Speaker 1>But not here in the US.

0:10:57.120 --> 0:10:59.360
<v Speaker 3>No, here in the HUMANS has been very different.

0:10:59.440 --> 0:11:02.960
<v Speaker 4>We just do not have a federal law that really

0:11:03.000 --> 0:11:06.960
<v Speaker 4>applies to what CLEARVIEWAI is doing. There are some states

0:11:07.040 --> 0:11:10.040
<v Speaker 4>that have privacy laws that give you the right to

0:11:10.320 --> 0:11:13.200
<v Speaker 4>access the information that a company holds on you. So

0:11:13.280 --> 0:11:15.960
<v Speaker 4>if you're in a place like California or Connecticut, or

0:11:15.960 --> 0:11:19.280
<v Speaker 4>Colorado or Virginia, you can actually go to Clearview dot

0:11:19.320 --> 0:11:23.000
<v Speaker 4>AI and request to know what they have in their

0:11:23.040 --> 0:11:27.040
<v Speaker 4>database about you and ask that it be deleted. But

0:11:27.080 --> 0:11:30.720
<v Speaker 4>the only state that has a very strong privacy law

0:11:30.880 --> 0:11:34.439
<v Speaker 4>that's directly kind of about facial recognition technology is Illinois.

0:11:34.880 --> 0:11:37.000
<v Speaker 4>And this law passed in two thousand and eight, the

0:11:37.120 --> 0:11:39.080
<v Speaker 4>rare law that moved faster.

0:11:38.840 --> 0:11:39.720
<v Speaker 3>Than the technology.

0:11:39.760 --> 0:11:42.240
<v Speaker 4>And I tell the history in the book, but it

0:11:42.320 --> 0:11:45.720
<v Speaker 4>says that if you want to use someone's biometric information,

0:11:45.800 --> 0:11:48.520
<v Speaker 4>including their face print, you need to get their consent

0:11:48.920 --> 0:11:52.199
<v Speaker 4>or pay up to five thousand dollars per use. And

0:11:52.280 --> 0:11:56.480
<v Speaker 4>so there has been litigation against CLEARVIEWAI and Illinois. And

0:11:56.720 --> 0:11:59.080
<v Speaker 4>you mentioned at the top of the program how Madison

0:11:59.120 --> 0:12:03.840
<v Speaker 4>Square Garden uses facial recognition technology to ban lawyers anyone

0:12:03.840 --> 0:12:05.880
<v Speaker 4>who works for a firm that sued the company. And

0:12:05.920 --> 0:12:07.720
<v Speaker 4>they do that in New York City at Madison Square

0:12:07.720 --> 0:12:10.240
<v Speaker 4>Garden and Beacon Theater and Radio City Music Hall.

0:12:10.640 --> 0:12:13.160
<v Speaker 3>But Madison Square Garden also owns a theater.

0:12:13.000 --> 0:12:16.680
<v Speaker 4>In Chicago, and they can't use the technology that way

0:12:16.720 --> 0:12:19.679
<v Speaker 4>there because they can't use lawyers, you know, faced prints

0:12:19.679 --> 0:12:21.160
<v Speaker 4>in Illinois without their consent.

0:12:23.080 --> 0:12:26.720
<v Speaker 2>How much when police enforcement here in the United States

0:12:27.120 --> 0:12:31.960
<v Speaker 2>have used this that it's led to mistakes or erroneous arrests,

0:12:32.000 --> 0:12:33.520
<v Speaker 2>if you will.

0:12:33.760 --> 0:12:37.400
<v Speaker 4>So there are hundreds of police departments that have used

0:12:37.440 --> 0:12:40.800
<v Speaker 4>clearview AI's technology. The Department of Homeland Security has a

0:12:40.840 --> 0:12:45.640
<v Speaker 4>contract with them. The FBI does I know of one

0:12:45.800 --> 0:12:48.280
<v Speaker 4>case where this has gone wrong, and there may be

0:12:48.360 --> 0:12:51.239
<v Speaker 4>others that haven't come to light, but in that situation.

0:12:51.480 --> 0:12:56.240
<v Speaker 4>A man named Randall Karan Reid was pulled over the

0:12:56.320 --> 0:13:00.320
<v Speaker 4>day after Thanksgiving in Atlanta, Georgia, and the there are

0:13:00.360 --> 0:13:01.119
<v Speaker 4>four police.

0:13:00.880 --> 0:13:01.839
<v Speaker 3>Cars that pulled him over.

0:13:02.000 --> 0:13:03.679
<v Speaker 4>They asked him to step out of his vehicle on

0:13:03.720 --> 0:13:06.719
<v Speaker 4>the side of the highway and started to arrest him

0:13:06.920 --> 0:13:10.079
<v Speaker 4>and told him there was a worn out in Jefferson

0:13:10.160 --> 0:13:13.880
<v Speaker 4>Parish for him for larceny. And he said, where's Jefferson

0:13:13.920 --> 0:13:16.600
<v Speaker 4>Parish And they said it's in Louisiana. He said, I've

0:13:16.640 --> 0:13:19.840
<v Speaker 4>never been to Louisiana in my life. And it turned

0:13:19.840 --> 0:13:23.119
<v Speaker 4>out that he had been identified through running a surveillance

0:13:23.160 --> 0:13:27.480
<v Speaker 4>still from a person who had been basically stealing purses

0:13:27.559 --> 0:13:30.240
<v Speaker 4>in and around New Orleans and it had hit on

0:13:30.440 --> 0:13:33.560
<v Speaker 4>Randall Reid, and when the police looked at his Facebook account,

0:13:33.840 --> 0:13:35.720
<v Speaker 4>they said, wow, he has a lot of friends in

0:13:35.760 --> 0:13:38.840
<v Speaker 4>New Orleans, and basically based on that, they issued this

0:13:38.880 --> 0:13:41.200
<v Speaker 4>warrant for his arrest. He was held in jail for

0:13:41.320 --> 0:13:44.920
<v Speaker 4>a week awaiting extradition until this was cleared up and

0:13:45.000 --> 0:13:47.480
<v Speaker 4>it became clear he wasn't the same person. He basically

0:13:47.559 --> 0:13:49.079
<v Speaker 4>took a lot of photos of his face and a

0:13:49.160 --> 0:13:50.960
<v Speaker 4>video of his face, and when the police looked at it,

0:13:51.000 --> 0:13:53.199
<v Speaker 4>they noticed he had a distinctive mole on his face

0:13:53.520 --> 0:13:55.240
<v Speaker 4>that the person they were seeking did not have.

0:13:57.120 --> 0:13:59.440
<v Speaker 1>Hey, Kashmir, Carol's died to get a question here. We

0:13:59.480 --> 0:14:03.679
<v Speaker 1>have a couple of minute stare each other very very

0:14:03.760 --> 0:14:06.400
<v Speaker 1>quickly talk a little bit about how oppressive governments have

0:14:06.559 --> 0:14:08.440
<v Speaker 1>used this type of technology around the world.

0:14:09.640 --> 0:14:13.000
<v Speaker 4>Yeah, so, China and Russia are a bit further ahead

0:14:13.000 --> 0:14:16.480
<v Speaker 4>of us in the deployment of facial recognition technology. And

0:14:16.640 --> 0:14:19.520
<v Speaker 4>for example, Russia has rolled out kind of real time

0:14:19.680 --> 0:14:23.760
<v Speaker 4>facial recognition on their surveillance cameras around Moscow and they've

0:14:23.800 --> 0:14:25.520
<v Speaker 4>been used to identify.

0:14:25.120 --> 0:14:27.239
<v Speaker 3>Protesters against the war in Ukraine.

0:14:27.920 --> 0:14:31.120
<v Speaker 4>I've heard about doppelgangers who get stopped on the street

0:14:31.120 --> 0:14:34.600
<v Speaker 4>because they look like somebody who's wanted. In China, it's

0:14:34.640 --> 0:14:37.520
<v Speaker 4>also been used to identify, you know, protesters in Hong Kong,

0:14:38.680 --> 0:14:42.840
<v Speaker 4>to to keep track of weaker Muslims there, and China's

0:14:42.840 --> 0:14:46.040
<v Speaker 4>also used it for kind of these these wild use cases,

0:14:46.160 --> 0:14:51.640
<v Speaker 4>like people who who wear pajamas in public in one

0:14:51.680 --> 0:14:55.480
<v Speaker 4>city were identified with facial recognition and named and shamed

0:14:55.520 --> 0:14:58.000
<v Speaker 4>on social media. If you jay walk there, you might

0:14:58.080 --> 0:15:00.320
<v Speaker 4>just automatically get a ticket.

0:15:02.200 --> 0:15:05.080
<v Speaker 2>One thing that stuck at struck me in your book.

0:15:05.080 --> 0:15:08.680
<v Speaker 2>The Chinese government has also started collecting DNA and tracking phones,

0:15:08.760 --> 0:15:11.800
<v Speaker 2>layering multiple systems in an attempt to create and all

0:15:11.840 --> 0:15:14.200
<v Speaker 2>seeing eye on its population of one point four billion.

0:15:14.320 --> 0:15:16.200
<v Speaker 2>We know that this is what China is doing when

0:15:16.240 --> 0:15:17.600
<v Speaker 2>you look at this and we've just got about a

0:15:17.600 --> 0:15:21.680
<v Speaker 2>minute left. More broadly, I mean, I do my fingerprint

0:15:21.720 --> 0:15:23.760
<v Speaker 2>multiple times a day to check in. I mean, we've

0:15:23.800 --> 0:15:27.160
<v Speaker 2>given up so much. What is the big concern about

0:15:27.200 --> 0:15:30.320
<v Speaker 2>kind of where this goes beyond? Is it all about police,

0:15:30.600 --> 0:15:34.200
<v Speaker 2>you know, arresting you falsely or what is it? And

0:15:34.240 --> 0:15:35.840
<v Speaker 2>again just got about forty seconds.

0:15:36.200 --> 0:15:37.920
<v Speaker 3>I love that I only have forty seconds. I would

0:15:37.960 --> 0:15:40.400
<v Speaker 3>just say, I there's many ways.

0:15:40.800 --> 0:15:44.000
<v Speaker 4>That's okay, that's many ways in which facial recognition technology

0:15:44.000 --> 0:15:46.480
<v Speaker 4>can be beneficial and that we want to use it

0:15:46.520 --> 0:15:48.760
<v Speaker 4>in our lives. But I think that we should have

0:15:48.840 --> 0:15:51.800
<v Speaker 4>the ability to take what we want and not be

0:15:51.880 --> 0:15:55.840
<v Speaker 4>forced to accept it just as being ubiquitous in our

0:15:55.840 --> 0:15:58.360
<v Speaker 4>lives that were tracked all the time, that we just

0:15:58.440 --> 0:16:02.920
<v Speaker 4>have no privacy. I think that's very dystopian and that

0:16:02.960 --> 0:16:05.440
<v Speaker 4>we should be able to choose the future that we

0:16:05.480 --> 0:16:08.920
<v Speaker 4>want and not just let what's possible with the technology

0:16:08.920 --> 0:16:09.440
<v Speaker 4>determine it.

0:16:11.360 --> 0:16:14.040
<v Speaker 2>Right safe to say we're living in a surveillance state already,

0:16:14.160 --> 0:16:17.160
<v Speaker 2>it's just how it's all used and where it goes

0:16:17.160 --> 0:16:19.400
<v Speaker 2>from here. What a fascinating book and so glad we

0:16:19.400 --> 0:16:20.120
<v Speaker 2>could have you with us.

0:16:20.200 --> 0:16:22.480
<v Speaker 1>Yeah, in YOURK Time Technology port Kashmir Hill. Her book

0:16:22.520 --> 0:16:24.600
<v Speaker 1>is called Your Face Belongs to Us, A Secret of Startups,

0:16:24.680 --> 0:16:27.000
<v Speaker 1>Quest to End Privacy as we know it. The book

0:16:27.040 --> 0:16:30.080
<v Speaker 1>is out now. Also, we went to graduate school together

0:16:30.240 --> 0:16:32.800
<v Speaker 1>and it's so cool to see all the success that

0:16:32.880 --> 0:16:35.480
<v Speaker 1>Kashmir has had. Congrats on just a great book, Kashmir

0:16:35.520 --> 0:16:38.400
<v Speaker 1>I really appreciate Kash taking the time to join us

0:16:38.400 --> 0:16:39.080
<v Speaker 1>this afternoon.