Facial Recognition Machinery, Part 1

What if your physical body included a searchable database of pretty much everything you ever did or posted on the Internet. So whether you're out at a bar with your friends, or you're sitting on the subway on your way to work, or you're sitting in your car in traffic, or taking part in a protest march, or working out of the gym, or you're on a date, whatever, anybody who could see you would instantly have the ability to look up your

personal information. They could find out your name, your resume, your contact info, your workplace, home address, maybe find all publicly available photos of you that are out there on Instagram or whatever. Maybe everything you've ever posted on Twitter, or on Facebook or whatever other social media, and more broadly, basically just everything you've done on the internet, purchasing history, search history, as well as your location history anywhere you've

physically taken your phone with GPS enabled. I think most of us would probably recoil in horror at the idea that we would ever lose the ability to be anonymous in a public place. But perhaps the horrifying part of imagining this scenario is that I think what I'm describing is not only fairly plausible in some preliminary ways, this

is already the case, at least in principle. All the foundations and support structure of this horrible hypothetical world are laid, and really all that's left to do is just kind of tighten the screws on it. And one thing we know about this world is that there there's no shortage of would be screw tighteners out there, Nope, especially if you can make some money by tightening screws, which you very often can. Uh So, yeah, we we should already know that there is very little privacy in the modern

technology sphere. Our phones, our social media accounts are advertiser i d s, which are used to track us across the Internet. These sources are already used to create profiles in which the disparate types of our personal identifying information get correlated with each other and used to serve us

ads or manipulate us on social media. But the leap into physical space, where all of our information is easily linked now to our physical body wherever we are, whether we like it or not, is the frontier that's currently being pushed and at a very rapid pace. Now there are multiple ways to make this link, of course, you know, so linking our digital profiles and all the the associated

data with our physical bodies. A very simple one would be with the tracking devices that pretty much all of us carry with us at all times, the mac addresses on our phones and our mobile devices. But one of the most powerful developing techniques is for the technosphere to recognize you in physical space the same way that your

friends and your family do by your face. Now, one thing to keep in mind about this is that, of course, just because say you're thermostat in your house can recognize your face, that is not necessarily in and of itself a bad thing. In some cases that could be very helpful, or even it could be seen as a way to uh, you know, to safeguard the temperature of your home, uh, that sort of thing, like you could have some sort

of a security feature. And as we as we proceed through these episodes, we're going to try and keep that in mind. We're gonna we're gonna try not to color the technology as inherently vile or inherently uh prone to misuse. But the story of technology is that it is both light and dark. Well, yeah, and that I think there is a big difference between something being inherently vile versus prone to misuse. There are a lot of things that are that are created with perfectly good intentions in mind.

You can think of lots of great reasons to do most of the worst technological stuff imaginable. You know, we've talked before on the show. We make no secret our distaste for a lot of things about social media. But of course, you know, you can see the good side of something like Facebook, Yeah, or very broadly, you can think of something like banking. Banking in many respects allows humans to do amazing things things they wouldn't or otherwise

be able to do. Uh, to buy larger uh, you know, to buy a property, you know, to buy a vehicle, to start a business, start a business, and so forth. But at the same time, very terrible things have been done and are still being done under the broad tent of banking. Yeah, and we can help better protect ourselves from those outcomes by better understanding banking so that we can regulate it properly, which doesn't always get done, but you know that that, at least in theory, that is

the way to protect yourself. And so today we're gonna be taking that same point of view to a series of episodes about facial recognition science and technology. And this is a subject I've wanted to talk about for a while because obviously this is an issue of increasing significance today. But actually, just after we landed on this topic, I came across a brand new story in The New York Times that serves as a really good anchor on why

this issue is incredibly relevant today. And this article is called The Secretive Company that Might end Privacy as we know it by cashmir Hill, published in the New York Times January. Yeah. It uh. It concerns a small technology company called clear View AI, which just the title of that it sounds like it could be fine, right, sounds very transparent, doesn't clear View AI, But of course this is a facial recognition based artificial intelligence company. Uh So,

so what is clear View in their own words? What do they say about themselves? Uh? To read from their marketing materials quote. Clear View is a new research tool used by law enforcement agencies to identify perpetrators and victims of crimes. Clear Views technology has helped law enforcement track down hundreds of at large criminals, including pedophiles, terrorists, and sex traffickers. It is also used to help exonerate the innocent and identify the victims of crimes, including child sex

abuse and financial fraud. Now, on the surface of things, that sounds absolutely air tight, right. It describes the technology that is used by the appropriate agencies to protect the innocent and to go after the guilty. But then again, that can be used to sell a lot of things in the world, of course. So what they advertise is that this app helps law enforcement identify perpetrators and protect victims of crime, and of course, in some cases that

may very well be true. Obviously, it would be pointless to deny that facial recognition technology the ability to take a picture of somebody's face and then find out tons of stuff about who they are and how you can find them. You know, be pointless to deny that in many cases that would be useful and beneficial to law enforcement. But it is also, of course just so easy to see how a tool like this could be terrible, both

in its successes and in its failures. So, first of all, of course, it could fail in catastrophic ways, say with like false matches when police are looking for a perpetrator, but of course you know that's something that can happen with human witnesses too, right, But then it could also be used effectively if it correctly identifies people too amazingly insidious ends. So how does it work. It's actually it

sounds pretty simple in terms of its user interface. Specifically, what this tool does is it matches an input photo of a face with a huge database of existing photos scraped from the internet, and then it will provide links to the places that those images from the internet were

originally found. So very simple example, I take a photo of you, and then I feed it into the tool, and then it comes back with other photos of you and links to the places where those photos were found, maybe your Facebook page, a YouTube video, you're in and so forth, and so when it works, it will provide a direct link between your anonymous face from a picture and the digital locations where all of your personal information

may be logged online. Now, we're not going to look too far under the hood of exactly how the you know, the underlying technology of this sort of thing works right now. Uh, there are a number of ways that facial recognition algorithms can actually work, but a very common way is using

neural networks. Yes, and uh and and for this For for like a nice succinct description of how this works, I'd like to refer to Max tag marks most recent book Life three point oh, which deals at length with AI and the potential the threats posed by by AI in the future. It's a it's a wonderful book. But in this this one section, he's just summarizing how this kind of facial recognition works. Uh. And he writes, uh that neural networks have been quote trained to input numbers

representing the probability that the image depicts various people. Here, each artificial neuron and on on an illustration they're depicted as circles. Computes a weighted sum of the number sent into it via connections or lines in this image from above, applies a simple function and passes the results downward each subsequent layer of computing. Higher level features. Typical face recognition network contain hundreds of thousands of urons. Uh. The figure

shows merely a handful for clarity, so uh. In the the visual representation the tech mark includes here you see all these circles and interconnected lines representing how the you know, the neural network is functioning. But then it begins, then we apply this to the facial features. It starts with with sort of general features and sort of blurry shapes, and then to more specific features, and then tying those features together and then eventually getting to an output probability

of actual facial matches. Yeah. And as with many other neural networks that are trained on large data sets, to you know, match values together or produce an output, why given input x uh you know, given a training method, there will often so. The way these things are trained is that you you know, you feed them a lot of examples of the kind of output you want, and

slowly they refine their own rules internally. The rules that happened at each of these layers of neurons to manipulate numbers and values as they pass through the neural network in order to give the output that closely matches whatever you've trained it to come up with. But that that means that like you can train potentially an effective neural network without yourself really understanding very well exactly what's happening

at each layer throughout throughout the network. Now, I think it is possible to like sort of get in there and try to dig into it and and see what's going on, if you've really got the time and expertise. But but it can be relatively opaque as far as computer programs go. It doesn't necessarily work like a normal computer program has lines of code that any programmer who knows the language can read through and figure out what's going on easily. But to come back to the cash

Mirror Hill article and UH clear view AI uh. One thing that's important to point out is that this is by no means the first facial recognition app or tool, nor is it the first used by law enforcement. It's particular value that the thing that it's doing that's somewhat new is in its database of images, which again have

been scraped from organic sources like Facebook and YouTube. Previously, law enforcement facial recognition matching programs were often weaker and more limited to smaller databases of government photos, say mug shots or driver's licenses. And of course there's the potential that, you know, smaller training or matching material will make any machine learning process weaker at at coming up with the

results you want. Yeah. All this kind of ties in with what I often think of its kind of like this title pool illusion of the Internet, that feeling that a lot of us had I think still have sometimes but also especially early on, this feeling that we were engaging in something segmented from the general population. You know. But the thing about tide pools, of course, is that eventually the tide rolls in and then you realize that

you're actually connected to the wider internet. Uh so you know, not just your friends or your family or your fandom, but also uh, you know, law enforcement, criminals, politics, all just churning around in the same grim ocean of numbing obscenity.

I think that's a really excellent metaphor. Yeah, there is some somehow the Internet was very easily able to create a sense of isolated, walled off gardens that we were living in which were at the time totally public you know, um, early days of various social networking sites, fan forums, all that kind of thing, you know, whatever it was that gave people a sense that they were in a little private space, you know, that their little corner, their little room.

But of course it's the Internet. What's happening there is public, and the consumers of what's happening there may be completely invisible to you. Right. So in this case, with the previous models of facial identification, the data sets they were depending on, we're basically title pools, like here's the title pool of of mug shots. Here's the driver's license, uh, title pool, and that's what we're feeding on. But basically clear View comes around, and this is a company that

is saying, well, let's just use the whole ocean. What's stopping is from using the whole ocean. So this is a company using the assets of various social media and in general visual media companies on the web to do the sorts of things that those companies have been loath to do, or if at least been you know, publicly opposed to doing. Because technically, as as pointed out in

Hills New York Times, article. Um, you know, there is there is an argument that what they're doing here, what what this company, in any company that's that's engaging in this kind of like broad sampling, that they may be violating the terms of service for these various websites. Sure, yeah,

automatically scraping imagery and data from Facebook. Say, I think there was at least the allegation that that could be a violation of Facebook's terms of service, but it didn't really seem to bother the clear view people right right, And I think when Hill reached out to Facebook representative, they said, well, we may look into that, and so

it's kind of an open question. But you know, a lot of this also comes down to something we discussed in our our look at during uh Lanier's ten arguments for deleting your social media accounts right now, because it concerns your data, data that you have, in all likelihood given away to companies like Facebook, Twitter and others simply

to be a part of the interconnectedness that they sold us. Now, a lot of the time when we're discussing such a data, we're discussing behavioral information, right your likes, your dislikes, your arousal patterns concerning posts and advertisements. But in addition to this, you also sold the devil your face. Yeah, I mean he he's the devil in this case is simply promising not to do anything unbecoming with your face. But but

Hell is highly populated. Even if you have good reason to trust this particular devil to which you've already entrusted your face and perhaps the faces of your family members, your loved ones, deceased loved ones, your children. Uh, you know, there are countless others that they will make no such promises, steal your face right off your head. Yeah, and and they may have little concern for, you know, the values

that were in place during the initial purchase. It's something that I think about coming up again and again with sharing data on the internet. So you share your data with a company, and you maybe trust that company today to protect your data. But what if so say that company hangs onto your data for a while and then they get new management that you don't trust as much, but they've got it, Uh, you can't get it back. Or maybe they have a security breach and somebody just

happens to steal your data from them. It's like, well, you would have trusted the company maybe, but now somebody else has got it, and you can see how in a world like that, Uh, it could start to feel like maybe hopeless or pointless or futile for people who say are in a position to make money off of not being very careful about people's privacy. Uh. You know, it's like, you know, what's the point. Everything eventually gets

out there anyway. And this kind of point of view was sort of articulated by some of the people quoted in Hill's article. For example, there's a figure named David Scalzo who was an early investor and investor in this company, Clear View AI uh, and Scalzo is quoted in the article saying, I've come to the conclusion that because information constantly increases, there's never going to be privacy. Laws have

to determine what's legal, but you can't ban technology. Sure that might lead to a dystopian future or something, but you can't ban it. So, you know, this is it at once one of those statements that seems very pragmatic but also entirely self serving, because true the story of technolo alogy is that its advanced cannot really be stopped. You have to think ahead as best we can and prepare our laws and our moral code to dealing with

emerging technologies. We've talked about this before, for instance, in as far as genetic technology is concerned, But this particular quote also sounds a lot like, Hey, it's gonna happen either way, so I might as well be the person to make some money off of it. It totally agree, I mean, I agree that it is difficult to stop

technological progress. That if you know, if one group of people isn't working on it, maybe a less ethical group of people might be somewhere, But that's not an excuse to be the person who creates the synthetic supervirus that you know who like genetically engineers, captain trips flu or whatever. Like. Also, you could use this logic about almost any bad thing. It's kind of like saying, yeah, you know, there's no

way to totally eliminate pollution. Some people are always going to find a way to pollute and harm the environment, so you might as well just go hog while just dump it all. Like. So, it's true that you can't stop everything harmful to the environment with regulation, but you can really slow it down. You can present major obstacles

to the worst types of offenses. And likewise, I think it would be very difficult to completely stop the advancing capabilities of AI, including facial recognition, but you can certainly slow it. You can certainly limit its potentially harmful uses by banning those uses and punishing offenders. Now, on the other hand, you could think, well, yeah, you could do that, but this would be so helpful to law enforcement in

some cases. You know, so would the ability to search any house you wanted without a warrant, right right, I mean, this is the same argument that has often been part of the reasoning for enhanced interrogation and torture, is that, well, it can help us get the bad guys, it can help us in this situation. And then also in all of these arguments are is also the the idea that well, if you have nothing to hide, if you are truly a good and supportive member of society, then what if

you have to what do you have to worry about? Anyway? But but it kind of comes down to the data issue. Well, you trust the person who has your data now, but do you trust the person who have your data tomorrow? You trust the government of today, but governments change, yeah, I mean, and No, nobody actually in practice believes this what do you have to hide argument? It's just something

you would say. I mean, like, if anybody ever says that, just immediately demand them to give you their email password. Me like, yeah, just let me read all your email. I mean, what's the problem, wild inspector? Right, I think

you'll find everything in order. Uh. But yeah, so, I mean obviously, society has often decide to regulate police power in ways that I mean that are truly inconvenient to law enforcement, because they decide that in some cases there are types of privacy and other civil liberties that matter more than prosecuting offenses at the maximum efficiency. Yeah. Alright, Well, on that note, we're going to take a quick break. Okay,

we're back. So again, we're in the middle of this first episode of this exploration we're doing of of facial recognition science and technology, and we've been talking about this New York Times article that just came out last week by Cashmere Hill about a facial recognition technology company called clear View AI. Now, broadly speaking, the two big advantages of clear View AI are that it first of all,

pulls from an extensive database of images. So we're talking three billion photos in a database versus the four hundred and eleven million searchable through the FBI's database. These stats according to clear View marketing materials reviewed by Cashmere here Hill for that New York Times article. And then secondly, it boasts a robust enough facial recognition engine built up from academic work by others on artificial intelligence, image recognition

and machine learning. UM that it and it and it it does not require high quality or complete facial images to produce matches. So like the you know, the I guess the ideal example would be you could have somebody going into a bank and robbing it with their face partially covered, and then this would potentially be able to match that partial face to a full face and say, a Facebook profile, at least according to what the company and some of its satisfied customers in law enforcement have

been alleging. Yeah. Like one example of a successful match that Hill mentions h is um matching an individual to a face in a mirror in someone else's gym photo. What yeah, and uh, you know, details of the you know, presumed guilt or innocence of that particular individual aside, I think this is notable in that gyms are often considered to be photo taboo places, and it's certainly as far as like other people working out in the gym go.

You know, this is I'm not, you know, an expert on jim etiquette, but it is my understanding that you shouldn't even you know, accidentally photographed someone else the gym. But obviously it does happen. Uh, it just it just you should not have your phone out snap and picks at the gym unless it's unless it's your private gym and you're the only person there, yeah, or unless it

catches bad guys, because what do you have to hide. Um. That being said, the people at the company do admit that of course, like you know it, it still has flaws.

There's still things that can't do. Yeah, Like, for instance, a lot of it is leaning on eye level photos, the kind of photos that you see and say a linked in profile photo, as opposed to the sort of ceiling level security camera footage that it is often involved in these scenarios, right right, uh yeah, I mean, And it's so it's running into the same kind of problems that we could talk about this later in the episode

that human beings sometimes have with less familiar faces. I mean, this is a A known thing about human face perception and facial recognition within the brain is that, uh, we are much better at recognizing very familiar faces under unfair ruble conditions, like a partial face, a face it a weird angle, facing bad lighting. We can do that a lot better if it's a familiar face. Then if it's a relatively unfamiliar face, right, I mean, even things like our face in a mirror versus our face in a photo.

You know, things like that can be distorting or if you or or more more directly, I find it with someone else's face reflected in a mirror. I'm definitely not used to seeing that, and that'll throw me off sometimes. Do you ever do the inversion test? This is another weird quirk of of the brain, trying to see if you recognize photos of people's heads upside down. I know we've talked about that before in the podcast, but I haven't really put it to the test in my own life.

There's a This is just a total side note. There's a very funny thing known as the Thatcher effect. It has to do with um, the fact that so if you look at somebody's head upside down, but with their eyes right side up. A lot of times people look at that and they don't even notice anything's wrong with the photo. So, like, the head is upside down, but the eyebrows are like over the eyes because the eyes

are still in the correct orientation. Um. But then if you flip that whole thing to where the head is right side up but the eyes are upside down, it looks unbelievably grotesque, Like you will burst out, you'll make noise when you see it. Look it up. That's your effect. But anyway, back to the story about clear View. So the company claims its product finds matches for an input photo up to seventy percent of the time, and of course Hill notes in the article that we can't be

sure how often false matches turn up. She quotes Claire Garvey, who is a researcher at Georgia University's Center on Privacy and Technology, who says, quote, we have no data to suggest this tool is accurate. The larger the database, the larger the risk of misidentification because of the Doppelganger effect. They're talking about a massive database of random people they've

found on the Internet. The Doppelganger effect being not that that vengeful German spirits are are actually invading the database, but that there's just a going to be the larger the pool of people, the more people they're gonna look very much like like others. There's gonna be more similarity

between an increasing a pool of individuals. Right. But at the same time, anecdotal reports from a number of law enforcement officers have claimed that this tool was effective at identifying real perpetrators from photos alone, and there have been plenty of other examples in recent years of supposedly effective facial recognition technology provided by other companies that have been used to UH to allegedly capture perpetrators of crimes done in public places in New York, in the UK, certainly

you know, in in countries with a very strong surveillance state, like in China. And we can come back to more about that in later episodes, I think, But as a personal anecdote in this reported story, Hill at one point has the company's founder use the app on a picture of her, and she claims that the tool quote returned numerous results dating back a decade, including photos of myself that I had never seen before. When I used my hand to cover my nose, and the bottom of my face.

The app still returned seven correct matches for me. So I think we can assume that failures, including both false and negatives and false positives are surely occurring at some rate. But it's also clear that this thing at least works some of the time. Yeah, and that's enough to help it get picked up by law enforcement. Also, it helps that there was a it seems like there's a pretty sizeable outreach campaign from the company to to to market

the technology to law enforcement. Yes, and and we should say, I mean, we're not going to hash out everything they get into in the article, but the company has arrived at law enforcement as their primary customers. Before that, they tried to market it in all kinds of ways, including you know, for like personal use to like private security things, and for like commercial use and even then like political

opposition research and stuff. Yeah, but this, uh, but you can definitely see the advantage to law enforcement here, because a detective, for instance, has their disposal a number a limited number of talents and tools that are useful and attempting to solve a case, and adding this to the to the toolkit is no brainer. Because you know, larger issues of stability of the platform aside, you know some

of these legal issues we potential legal issues we're discussing earlier. Um, you know, this would be something you could use in congress with other techniques. You know, you could say, all right, this face seems to match up with this individual. We also know that this individual was in the correct you know, vicinity at the time. You know, you could lean on your other detective tools then to to actually make the case. That's not to say there's not potential for misuse here,

but uh, I'm just saying you can. You can definitely see the appeal and how if everything is working perfectly, it would be an effective law and forcement tool. Yeah, and however effective it actually is, it's clear that this and similar tools are increasingly popular with law enforcement in countries all over the world. If you're one of those people who feels like pumping the brakes on this kind

of technology, what could actually be done about it? Well, Hill quote somebody named al Ghadari, a privacy professor at Stanford Law School who says, bluntly quote, absent of very strong federal privacy law, we're all screwed. Uh. And He's not alone. There are plenty of privacy experts today advocating the point of view that facial recognition technology, or at least some specific uses of it, aren't just something that maybe we should be a little concerned about, there's something

that needs to be banned outright. Um. For example, Hill also quotes somebody named Woodrow Hertzog, who is a professor of law and computer science at Northeastern University. Uh and hart Sog says, quote, we've relied on industry efforts to self police and not embrace such a risky technology, but now those damns are breaking because there's so much money on the table. I don't see a future where we harness the benefits of face recognition technology without the crippling

abuse of the surveillance that comes with it. The only way to stop it is to ban it. So whether we should do that, or if so, what form that band should take, what whatever, is the best way to address it. I think it is at least clear that this is a very pressing and uh like time sensitive issue,

that that is of urgent public concern right now. Yeah, because, again, as the author points out, I don't think any of us want to live in a world where any stranger can you know, surreptitiously take a photo of our face and then face searches and get all this data on us. You know, I don't want for us to build that kind of world for our children, who, more than any of us, never had a chance to opt out of

this uh, this face trade, you know. And that's the that's the pressing to think about, you know, because because you're not thinking about that when you when you share images of your child on on Facebook or Instagram or whatever it happens to be. Um, you know, you you just you're wanting to celebrate that this person exists at all, but you're you're laying the groundwork from like you know, age zero Hola onward right, Uh, like this is their

their digital history. We were talking before we came into the studio about like if a person wanted to do something about this, what could you do? You can't unpost photos of yourself and data that has already been scraped. But I wonder if maybe you could try to come up the works by constantly just polluting the Internet with false pictures of you that are not you. So you're like sort of like deep facing enough to where you um,

you've obscured the visual record of yourself. Maybe yeah, like if you, I guess it would then depend on the all those new images being taken in and causing enough confusion and the identification of you. But I don't know, or perhaps like altering your facial appearance with enough regularity that there is no concise version of you, or or at least making it to wear. Then the the AI

would have to work a lot harder. It would have to have like a broader definition of what you look like, to the point that maybe it excuse with maybe it enhances the doppelgang or effect, like I'm thinking about you know, like you just you know, each day you inject a different portion of your face with collagen or something, or maybe not collagen, but maybe just saying lead, well, I mean, does this lead to if you, I mean, this sounds ridiculous,

but does this lead to a future where everybody starts walking around with a broad array of interchanging masks? Yeah, and will and then you have an enhanced um laws against the wearing of masks. I mean, masks are outlat in a lot of places and and a lot of events for for a reason. Yeah, you're not supposed to drive a car wearing a mask. That's um, this does remind me. There's a there's an excellent show on Hulu, a titled Future Man that is a it's it's a it's a comedy. It's a satire with a lot of

nostalgia for various you know, sci fi franchises. But there is a scene in one of the episodes where an individual knows that a facial recognition system is looking for him, so he gets himself beat up first so that his face is then all swelled and distorted and it and then it cannot make a match for him, and he's able to sneak past the guards. I don't think that's

a sustainable strategy. It's not. And more to the point, yeah, we should not even we should not have to even entertain that possibility to to hold on to, uh, you know, our sense of privacy. Yeah. Now, the fact that we're talking about this story in the New York Times about

clear View is it's just a result of timing. It's like this one specific company is not the entirety of the end of privacy problem, nor of the facial recognition technology landscape in particular, Another company could do the same thing. Other copycats I'm sure are already getting in there. Uh, it's just one high profile example of of the potential already being put to use that that's getting a lot of attention in the past couple of weeks. Yeah, in part two. You have to read the full article for

the details. But it's also like their key individuals that are notable that are tied into its funding. Yes, there's that, and and of course there's the ominous way it ends, which is the idea that it will soon probably be rolled out not just to law enforcement, but to be a publicly usable app, you know, which I guess is the sort of the scenario we were describing at the beginning of the episode, just having a publicly available personal search bar tool. Yeah. Well, well, markin me down for

being against that. Yes, alright, time to take a quick break, but we'll be right back with more than thank alright, we're back. So, of course, I also want to make the case that facial recognition technology is not necessarily always a dangerous or scary or ominous thing. I mean, I think there are some uses of it that one could

quite easily find benevolent or even delightful. And one of the ways I was thinking about this was many of its developing uses in non human animals, not all because some of its uses in non human animals are also like kind of horrifying, but some of the ones that

non human animals are pretty great. I was reading an article in New York Magazine from October called Here's a list of every animal humans currently monitor using facial recognition technology by Mac dagaron As A complete list is probably wildly out of date at this point because this was, but a few of the entries include things like there's a Norwegian fish farming company called Sermac Group as that commissioned a system for facial recognition of salmon which would

use distinctive patterns of spots around the eyes, mouth, and gills of individual salmon to build individual digital medical records associated with each fish. And this would be for the purpose of fighting epidemics of parasitic sea lice primarily. Okay, so not merely just presenting it with with the dish when you ordered it a restaurant, but I will take the baked salmon and please include it's complete medical history. Yes, this salmon was named Jeffrey. Here's his Facebook profile, you

know on on salmon Facebook. Oh, but that comes back. Of course, facial recognition technology is being deployed to keep individual track of all kinds of livestock like cows and chickens for maybe medical reasons or reasons having to do in the case of cows, reasons having to do with tracking, like periods of peak milk output and stuff like that. But there are also stories about conservation efforts to non invasively monitor wild populations of vulnerable animals by way of

facial recognition, which if that works, that sounds awesome. Like I was looking at article in Scientific American that described efforts to use facial recognition to track wild lions through a platform called the Lion Identification Network of Collaborators or link. That's interesting. It reminds me of I want to say, like a decade ago, maybe a little a little further

back in time. Uh, there was a piece I read about tracking whale sharks, and whale sharks all have you know, distinctive patterns on their you know that's sort of the top of their heads that area, and uh, you know, it doesn't mean anything, you know, much to humanize. But I think at the time they were utilizing NASA technology that was aimed at making sense of of the stars like astronomical um computation systems to then to make sense and sort of track um identify at any rate these

various whale sharks. So this would but this sort of thing would be I think an even better method of doing that because because you're you're probably dealing with it with creatures in all these cases that uh, you know,

there's a they definitely are not all identical. There are differences, but we just may not have the eye for it, whereas technology can be can be used to say the Sharper eye for chicken identity exactly, and you know, it has the advantage of not having to physically tag the animal in some way, which can be difficult to do or it can be harmful to the animal right or

day are dangerous to the individuals doing the tagging. Yeah, um and so apparently so you've got this one with lions, the link project, But there are similar things that have been attempted with tigers, elephants, even whales. You mentioned whale sharks, but with like actual mammal whales, with a project that an article in the Atlantic called Facebook for whales. I hope it's not as addictive for whales as it is for humans. That was a joke, but you didn't laugh.

That's okay. But similar technologies have been proposed and tested to help link people with lost pets, including cats and dogs. That seems like a great use of this. Yeah, like, you know, certainly, I'm on enough social media boards where their pet owners and occasionally pedicle missing. And then there's this whole back and forth where like like, oh my my orange cat is missing that it looks like this, and then somebody would be like, well does he look

like this? And I think I saw him in the backyard, and someone else is like, oh, I think I saw him over here across town, and nobody can be for sure, right, because it's hard to get close to a stray cat in some cases, or or an escaped cat or a feral cat that is then misidentified as a lost cat. But if you had the ability, if you had some sort of app infrastructure where your your cats missing, fine, you upload them to this database and then when someone finds a cat, they just take a picture of it

and it tells you if that cat is missing. Like that would be that would be great. That would cut out a lot of the the the anguish and the work that goes with having a runaway pet. I agree that sounds great, and maybe I'm suffering from a lack of imagination, but I'm I'm thinking in cases like that, that's a case where I think the the risks to the cats privacy would be far outweighed by the benefits of people finding their lost animals, because we know how

cats are. They don't give a damn about privacy. Now they have, Yeah, they have a whole different set of set of values. Now. Another sort of quirk of timing here as we were putting this episode together, in fact, is we're sort of finishing our no for this episode. This morning, I actually read a new blog post, uh titled depth of Field Fails by Janelle Shane at ai

weirdness dot Com. Oh, we've talked about her blog on the show before because it came up in the pair of episodes we did called flat a Sex Mackinah, which

was about why it's so funny when machines fail. Yes, yeah, yeah, this is and I imagine a lot of you have encountered, uh, you know, the various scenarios she's run with, like AI s coming up with names for Halloween costumes or names for I think at one point evenel like She's they were using it to come up with not only names, but actual recipes for cocktails, yes, recipes for foods, also names for dishes. We talked about D and D. Character

biography is generated, and and spells names for spells. Remember remember song of the Darn Daving Fire. Yeah, there's even one about cat names. But in this particular post, um Shane the research scientists. She tests out the facial recognition AI that is employed by Skype for its blur My Background for All Calls feature. Now, the curious thing about

this is I've been using Skype. We've been using Skype here on the show for years for interviews, and I guess I just I don't dig in deeply enough or read emails because I didn't realize this was a feature at all until today. I think we always let somebody smarter than us figure out how to use Skype and

there we'll just do the talking. But but it, but it totally makes sense as a feature because perhaps you do have an ideal environment set up for a call with a business friendly background behind you, but maybe you don't.

Maybe you have a fridge with a bunch of notes stuck to it with magnets, and maybe some of those are like bills or you know, or they have some data on there that you you wouldn't even want there to be a chance somebody might be able to decipher it, or a bookshelf full of occult tones that you don't want people to know you've been researching that that's a

good one. Or perhaps you're at work and there's a marker board full of you know, of data back there might be something you don't want out, Or perhaps you have some distracting art up there in the wall and you don't want to compromise the interior of your own home so that you can do a skype call. You don't want to have to like take things off the wall in order to do this, or you're in one of those weird Roman mass toilets, whatever the case may be.

The AI then can auto blur all of that out for you, But to do so, it has to be able to tell the difference between the face of the collar and mere objects in the background. Uh. And it's in this case the AIS definition of a face is pretty broad. As Shane discovers, it will allow ancient Egyptian illustrations to to come through unblurred, to be the face, to be the face. So it's like, okay, this call is being made by this individual from ancient Egypt isis

but also increasingly abstract depictions of a human face. Uh. So it showed various works of art and it's some of them were really abstract, and it was like, all right, that's a face. Sure that will work, just like monk the scream, Yes, yeah exactly. Um. Also stuffed giraffes. It had a problem with the like the horns of the draffe, but but not so much the face of the stuffed giraffe. Uh. It gets a little confused with life size plastic skeletons, however,

and also cats can throw it off as well. But so do check out that that blog post. It's it's amusing and also insightful. But all this is certainly I think an example of facial recognition AI doing something that's not only helpful but could actually help you with your privacy. Now, one thing I think that would be different there is that that's facial recognition in the sense of recognizing a face as opposed to a background, versus recognizing whose face

a picture is of right. But then again you could easily imagine like that being an upgrade or being a situation where if you had a more robust um facial recognition um AI that was then used on some sort of skypelike system, you might actually go ahead and have a feature where the caller's face was logged and therefore it would blur out any face that was not the authorized users face. So that way of you know, there are other employees walking by in the background getting coffee,

they're not going to show up on your call. If you're you know, significant other walks by in the background, they're not going to show up on the call, etcetera. So even with this, I mean, when I'm in these kind of scenarios in my head, I'm always wondering if there are freaky applications that I'm just not being imaginative enough to to get to. Yet, We'll let me put on my black mirror neural lace cap for a second and think, um how about a simple case where law

enforcement wants to access an unblurred background. Now, I'm not sure to what extent that's even possible with this technology, but what if, say, you know, a government agency made a claim for a need to override the auto blurring features utilized by others, so they would just have blanket

power to do this. So you think you're blurring your background, but actually you're Yeah, nobody can see, not if you're on the phone with you know, with with with someone who's actually a government employee or whoever happens to have the magic key in this scenario. Oh, I guess it's kind of like how you think you can have your phone turned off, or you think you can have GPS turned off, but in fact it is still location tech. Yeah, yeah, that's sort of thing. And again I don't have a

detailed enough knowledge of this particular software. I'm not saying not not not applying this directly to the Skype scenario here, but just sort of thinking in general. Um, now, in this particular case, I'm also assuming that the broad definition of a face is in place, at least in part to avoid situations where a human beings face is blurred

because the AI can't handle, say, facial disfiguration. Because I think we can understand why we wouldn't want an AI like this to lean heavily on norms to promote ideals about who and who doesn't have a face? Sure though this in considering this with face recognition, and we get into some some interesting and you know, at times disturbing territory.

Matthew Galt had an article in Vice last February titled Facial recognition software regularly miss genders trans people, detailing how these systems were simply not built with trans or non binary people in mind, and can quote continue to reinforce existing biases? Oh yeah, I mean, as with a lot of things, I think sometimes there is an illusion that machines somehow will be free from applying biases to humans

that that humans apply to each other. But I think we we've got ample evidence now that that is not the case. That human biases get quite easily mapped onto artificial intelligence through assumptions used in the in the creation of these algorithms, or through the data sets they're trained on.

Right and then, and as Galt explores in the article, like part of it too is just like who's building these programs, You know, it's built being built by programmers and engineers, people that are that that may just not have have ever really given serious study to some of like say the gender issues that are you know inherent

to the problem. He also mentions how past databases have, for instance, misidentified black people in criminal databases and even in some cases had they failed to see black people at all. Yeah, like say, if they are trained primarily

on data sets with lighter skinned faces. In fact, just last December December of twenty nineteen, the National Institute of Standards and Technology in the US tested a hundred and eighty nine facial recognition algorithms from ninety nine developers, which includes like some big name developers, and found that they were far less accurate at identifying African American and Asian faces compared to Caucasian faces, and African American females were

even more likely to be misidentified. Uh. Now this was reported in various places. Uh, but the article I was reading about it BBC News text article facial recognition fails on race Government study says Yeah, I've read about several cases like this. Uh. I mean, I think it's just so important for people, especially working in the technology space, to remember, you know, don't fall for the myth that it's unbiased just because it's a machine and not a person.

People's biases end up in the machine. The rules come from us. Okay, well, I think we're gonna have to call the first episode right there. But when we come back in the in the next in the series of episodes, we're gonna be definitely talking about facial recognition in the organic brain, and we'll be moving on more to the history of technological facial recognition. We'll get to talk about Greebel's. We love Greebel's. Greebles were new to me, but there's

a whole world. We could do a whole podcast on Greebel's. You might think they were new to you, they weren't new to you. We've talked about Greebeles, we have talked about Greebles have the most delectable spikes. Okay, if you're curious, you'll have to come back next time to find out. Come back for the Greebel's. In the meantime, if you want to check out other episodes of Stuff to Blow your Mind, you can find us anywhere you find podcasts.

If you want to just a handy way to check us out, go to stuff to Blow your Mind dot com and that'll shoot you over to the I Heart listing for our program. But wherever you get the show, just make sure that you subscribe, make sure that you rate and review. These are great ways to help us out and just tell a friend about the show. That also helps. And don't forget our other podcast, Invention. Invention

is a journey through human techno history. Oh yeah, I feel like we've been doing a lot most of our technology stuff on invention these days, and so uh so I'm glad to be getting back into the techno space a little bit on Stuff to Blow your Mind today. Yeah, absolutely, even if it is for a kind of dystopian sci fi topic like this. Anyway, huge things as always to

our excellent audio producer Seth Nicholas Johnson. If you would like to get in touch with us with feedback on this episode or any other, to suggest a topic for the future, or just to say hello, you can email us at contact. That's Stuff to Blow your Mind dot com. Stuff to Blow Your Mind is a production of iHeart Radio's How Stuff Works. For more podcasts from my heart Radio, is that the iHeart Radio app, Apple Podcasts, or wherever you listen to your favorite shows.