Us. The reason why im excited about today is cashmere has been a pioneering kind, trailblazing voice in surfacing these issues of the social impacts of technology. What were always focused on at all tech is human, is really this gulf between how fast innovation moves and how slow our ability to consider its impact on how we live, love, learn, even die, how we get jobs, how we see the world, the future of the human condition. These are big questions, and thats why these are questions we cannot deal with sitting alone in front of a screen, come together because we need all different backgrounds, all different voices. But going back kashmir hill and just a very role that journalists play, journalists are informing about what is happening these oftentimes things done in a secretive manner. So especially around her work that came out with big New York Times expose a year or so ago around clearview ai oftentimes what one founder might want to do is not in a line with our interpretation of the tech future that we want and so much of what were trying to do here at all. Tech is human is cocreate a tech future aligned with the Public Interest in line with our values. And thats why it behooves us to make that a small sliver of society does not dictate the future of our tech, which is the future of our democracy and the future of the human condition. So people like kashmir hill are really exposing whats happening. So when we think about a topic like facial recognition, one of the concerns i think a lot of us face and have in surface is that it feels like its being placed upon us without our implied consent. Right, that the technology is released into the wild without us having some modicum of control of consent over how its designed, develop and deployed. And this is the part that really needs to change. And this is why and ill take a human we try to bring together stakeholders across Civil Society government, industry, academia. Again, the Important Role of journalists who inform the general public, who influence policymakers, who influence tech companies. Theyre all parts of the puzzle. So when you look around today and after todays discussion, when were all mingling and if you still dont want to leave after we close down here at 830, well continue the conversation next door at brass monkey. Is this good to meet other people who deeply care about where this is headed . Because the real important part is that our future is not decided our future is being determine end by what we do or do not do today. So without further ado, i have the pleasure of welcoming kashmir hill and our executive director of articles human rebekah tweed to be in conversation. Please give them a warm welcome. Thank you, david and thanks everyone for coming. Tonights very excited to be here in conversation with kashmir hill hill. Kashmir is a tech reporter at the New York Times and the author of your face belongs us, which you can pick after this. And kashmir will be signing for you over there. She writes about the unexpected and sometimes ominous Ways Technology is changing our lives, particularly when it comes to privacy. She joined the times in 2019 after having worked at Gizmodo Media Group fusion Forbes Magazine and above the law, her writing has appeared in the new yorker and the washington post. She has degrees from Duke University and new york university. She studied journalism. So welcome, kashmir. Thank you so much. Its great to be here. You are. Yeah. Glad youre here in new york. Youve been on a whirlwind book tour and a lot of the book actually takes place here in new york. So can we set the stage . Can you just give a little bit of an overview . You start working at the times, its 2019. Someone approaches you with a tip about a startup. They have scraped billions of photos, faces and apparently they can determine who a person is just based on a photo alone. So can you tell us what this company is . And what about this company is different and whats concerning to you . Yeah. So i find out from a records researcher that police are kind of secretly using a Company Called clearview ai, which at the time had 3 billion photos scraped from the public web. And there was very Little Information out there about the company and i remember on their web site, they had an address here in manhattan, it was just a few blocks away from the New York Times building. When i mapped it on google maps. But then when i walked over to it. The building wasnt there, it didnt exist. And it was the first of many kind of weird flags about investigating the company. It wasnt obvious who was working for them and there just wasnt a lot out there as i was trying to dig into them. It was clear that this company that was exposing so much about us was trying to stay in the shadows and what they had done with was was really shocking. We just hadnt the us seen Something Like this done before and it was this company that no one had ever heard of. Yeah. And can you tell us a little bit about who is behind company . Can you tell us about who and on top . Yeah. So the main person i think of as being behind clearview ai is a young guy named quinton tat. He grew up in australia with obsessed with computers from an early age at 19 years old he dropped out of his college where he was studying Computer Science because he thought what the professors were teaching was kind of boring and moved to san kind of chasing the dream of Silicon Valley and was basically just trying to make it there for a while. It was 2007. He was he made facebook quizzes. He made iphone games, kind of stuff at the wall to see what would stick, but nothing did. And you know he ended up moving to new york 2015, falling in with a very kind of conservative of crowd and the people that would become his cofounders, the company that became clearview ai. Yeah, something i thought was really interesting about his story. He doesnt have like some, you know mission that hes chasing. Definitely. He just wants know whats going to be successful, whats to stick. So these facebook quizzes this you know he does like an iphone app like trump hair and hes really just seen what you know whats going to work until it seems like once he hits facial recognition something is different there he really kind of goes down this like dark rabbit hole and he you know, you outline in the book this kind of like long history of people thinking they can determine about a person based on their features. So it was just really interesting to me that he seemed pretty obsessed with can you determine who is a shoplifter or a border crosser, someone whos crossing illegally or someone whos a terrorist or, you know, a cheater so im really curious if you can talk little bit about, you know, some of the use he was thinking of when he stumbles onto idea for facial Recognition Technology, like how did he get there . And what was he trying to do with it before he discovered the use case . Law enforcement, yeah. I mean, set the scene. So 2015, 2016 this time that one contact comes to new york is a really big time artificial intelligence. Its when theyre really starting to make advances in computer and Machine Learning or neural Net Technology and this is what powers so many of advances that surround us now kind of chat, you know, image generation. And it was the same thing with facial recognition was a time when computers were just getting the and the Software Techniques were just getting. So much better for computers to process all this data to recognize patterns. And so went to in his the people he kind of founded the company with originally were thinking you know can we take a photo of somebody and just from their face alone understand more about them. They werent originally thinking about just recognizing a face, putting a name to a face, but what about their what can we learn about them from their facial features . And they were tapping into the ideas that we we consider are pretty outdated about what you can just tell a person from their face. And so one thing he talked about doing with his cofounders was this was around the time that ashley was breached. Does everyone what Ashley Madison is so extramarital dating site and so a place to go if you want to cheat on your partner and the people hacked the site exposed all the users along with their names and their addresses is and email addresses. So he oh we should take this of people look them on facebook download all their faces is and then we can train a computer know what a cheating face looks like they had same idea about intelligence that you could figure out you know give it a whole bunch of photos of people with high iq and figure out whos most intelligence. Same thing with criminality they really had this idea that computers could kind of data mine who we were and figure that out from your face that their business did not end up being very successful and huan contact says now you kind of renounces idea that computers would be capable of that and they went instead to something that was much easier by comparison which is which was just download a whole bunch of photos from the internet and make this tool that when you upload a photo of somebody it shows you all the other places where their face appears along with a link to the website and so thats where you can find out their you know their social media profiles who they know maybe photos they dont even know are on the internet. Yeah. And i think its really to see how that develops into the this use case for Law Enforcement agencies and thats a really interesting story. Can you tell us a little bit how he discovered that that was where he wanted to. That was the target market for, his product. Yeah. So clearview ai originally was called smart checker and when they first started doing this one of the first places actually that they got access to photos was venmo dot, which at the time you know, venmo makes everything public, died by default. And on venmo dcoms web site, they had real time transactions that were happening on the network like people paying people and it would have their photos in a link to their Profile Photo and they ended up downloading all of photos. And so originally smart checker was just like a way to find take a picture of somebody and find their venmo account. But originally theyre just trying to figure out who they are, who pay for this. Like who would this tool for taking a photo of somebody and find out who they were . And so they pitched it to hotels and companies as banks, you know, figure out who the High Net Worth people are as they walk in. I mean, theyre really just kind of giving it to investors, like giving it to Grocery Stores and had all of these kind of the first people to use clearview ai were essentially very wealthy people, many of them here in new york. My favorite story is john catsimatidis, who is resident billionaire, owns the christie, so i always pronounce it incorrectly. Is it christie . Chris christie . These markets, they pitched it to him to have in his Grocery Stores, to identify shoplifters. He told me he was having a really hard time with haagendazs thieves like people were stealing ice cream from the stores but they also just gave the app to him to put on his phone and i was like, well, how do you use . And he said, well, one time i was having at cyprian, cyprian is the italian in soho and my daughter walked in and she had this date on her arm and i wanted to know who he was. So i had a waiter go over and take a photo of them. And then i ran the photo through clearview ai and i figured out who the guy was and i approved, but it was really they were talking to a real estate building about maybe putting clearview ai to kind of check the identities of people coming into the building and the security chief there was vetting the app and he used to work for the nyp and he said, wow like, this works very well and this would be a tool that my colleagues, my former colleagues would love. And so he introduced them to the nypd. So its almost happenstance that they ended up being used by police. And then it was very popular with the nypd. Yeah. And you can see how people have a tendency to want to misuse a tool like that for their own personal uses and it seems like that part of what was happening with potentially with Police Departments as well where having to where the nypd wanted to to build in some checks to make sure that didnt happen during a mention anything were on that yeah so the nypd originally at first was used by the financial crimes unit there and you know they tend to have a lot of photos of fraudsters as an atm at a bank counter and so they started running these photos through clearview ai networked well and they told their colleagues and so clearview ai just offering these 30 day free trials to any police officer. So they just handing it out like candy to the nypd and its really striking to me because i dont know. I first started looking into this. I didnt realize police could just use any tool they wanted like just from some random guy download it and actively using it in investigation ends without kind of checking the accuracy of the algorithm, seeing how well it works. It was really just yeah. Just, just try it and see you like it. So it starts spreading among the nypd they start telling the people they know at other Police Departments, you know, the department of homeland security, it just kind of like started spreading like wildfire or through Law Enforcement as as one investor put it. And the one thing the nypd was worried about is, you know, if we start giving it to all of our officers, how do we make sure that . Theyre only using it for purposes and not theyre like out at a bar and they see a, you know, pretty person and run his or her photo and find out who they are and they were starting to ask clearview, like, how how do we kind of keep track of how many searches are being run . How do we make sure theres kind of a Case Associated with it and clearview ai was kind of building in kinds of ways of monitor ring usage in reaction to the police as opposed to before started using it. Yeah, thats so interesting. And one of the other things i find really interesting, the book kind of bringing it back to. The one that theyre is a time where you you really outline his foray into maga world basically and how he starts associating with a lot of you know some far right personalities like Chuck Johnson and and also the deplorable of played into the development of the tool you know smart checker at the time that became clearview. So could you talk a little bit about the impact of ones proximity maga world . Yeah, some of the early use cases of the company the tool that became clearview ai were pretty alarming to me. And so one of the basically the first time it was used, it was still called smart checker at the time was at the deplorable in washington, d. C. , which was event around trumps inauguration to kind of celebrate all the work that people had done to help get him elected. And they want to make sure they didnt have anybody from, the far left coming in. They didnt want any antifa people there. And so apparently they kind of searched the people who bought tickets and they claimed that they identified two people who were affiliated with the Antifascist Coalition in d. C. And made sure that they didnt get tickets and i found out about that because they used that use case when they were pitching technology to the hungarian government as a Border Control technology. And they said, you know, we weve actually they they claimed in this presentation that they had find tuned their product, identify people with george soros and the open Society Foundation who are kind of pushing for democratic reforms in hungary. And they said here this would be a great tool for keeping them out. And so just in, i think its because there were really affiliated with these kind of conservative causes, but it shows how how you could use a Technology Like this in a very chilling political way. Yeah. And one of the things i think so interesting when you were in the book about whats different about clearview you are not necessarily a technology breakthrough, but an ethical breakthrough of sorts where they were willing to do what other others were not. And you talk about how scrapes, billions of photos of faces using basically what was available. And there you know, the Ashley Madison example is classic and then venmo that thats so interesting they basically took this real time feed were going to just like pull these photos but then facebook you know making Profile Photos public by default and then being able to to scrape all of those photos. So facebook as assures people that the technology exists to make sure that you cant scrape those photos. So what i was confused about when i was reading it was how did clearview come through and end up scraping of those facebook photos . They werent supposed to be able to. Yeah so there is this theres this problem, all of us in terms of dealing for privacy is that it is . One hard for us to understand whats going to happen in technology that we need to protect against. And so i think many of us who have posted photos of ourselves and others and loved ones on the internet over the last, you know, two or three decades or anticipo stating that a company like clearview ai was going to come along and scrape all and, you know, organize the internet around our faces. So, one, its hard for us to to kind of predict that and ourselves against that. And then this other question of what kind of Technology Companies have done or not done to protect us. And so with facebook, i mean, i, i think facebook such an Interesting Company because it forced all of us t