In short
The episode covers two stories. First, hackers linked to Shiny Hunters claim they “hacked the FBI” and obtained data from an FBI jobs portal, including 5,000 FBI officials and their spouses (and sometimes parents), with names, addresses, phone numbers, and job roles. Reported sensitive contents include urine/blood tests and psychiatric evaluations. The data allegedly includes people tied to the FBI’s Remote Operations Unit (its hacking team) and agents investigating Russia and China. Shiny Hunters previously used similar ransomware/doxxing tactics (threaten leaks unless paid) and has been linked to ICE official doxing. The FBI says it’s investigating; Shiny Hunters demanded the FBI correct a cybersecurity report and claimed a one-week countdown, later saying it was a “marketing campaign” and they wouldn’t publish.
Guests
Sam Cole and Jason Kebler, 404 Media co-founders. Second story: Jason reports Vidizmo wants to add facial recognition to Flock cameras via Flock OS integration, despite Flock’s stated refusal; Vidizmo pitches searching Flock + Axon evidence by face/vehicle/object and offers features like race/gender/age analysis.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOOverview of the FBI Hacking Incident
1:00 to 2:48
Discussion of the recent hacking incident involving the FBI and the hackers known as Shiny Hunters.
“I'm your host, Joseph, and with me are two of the other 404 Media co-founders.”
Who are the Shiny Hunters?
2:48 to 5:16
Exploration of the Shiny Hunters hacking group and their methods.
“And this isn't just like some rando hacking group or a random person.”
Details of the Data Breach
5:16 to 9:18
Investigation into the types of data that were compromised in the FBI hack.
“They are, frankly, a prolific hacking group, collective, amalgamation.”
Motivation Behind the Hack
9:18 to 14:01
Discussion of the reasons why Shiny Hunters targeted the FBI and their demands.
“Did they say at this point why they did it?”
FBI Data Breach Analysis
14:01 to 18:02
Discussing the implications of the FBI data breach and the potential risks involved.
“just like they've been doing here with shiny hunters and the FBI data.”
Facial Recognition and Flock Cameras
20:56 to 28:00
Exploring the implications of adding facial recognition to Flock cameras.
“We have to take a quick break because Jason's mouse had died because of the amazing design of the Apple mouse.”
Facial Recognition and Police Surveillance
28:00 to 29:03
Explore how police can utilize facial recognition technology with surveillance data.
“cops can export that data and then they could just manually put it into Vidismo system or any other facial recognition system and run facial recognition on it.”
Vidizmo's Capabilities and Implementation
29:04 to 30:37
Learn about Vidizmo's offerings and the implications of their technology.
“itself from Flock in some way, either by offering ALPR with live video, offering it with more privacy than Flock offers.”
Controversies Surrounding Facial Recognition
30:38 to 32:51
Discuss the accuracy and ethical concerns associated with facial recognition tech.
“which is pretty common for facial recognition systems like Clearview.”
Interview Insights with Vidizmo's CEO
32:52 to 35:38
Gain insights from an interview with Vidizmo's CEO regarding their technology.
“So you can go into the system and say, show me all of the Latino males age 30 to 50, and it will pull up footage of that.”
Show all 12 chapters
The Business Implications for Flock and Vidizmo
35:39 to 37:44
Understand the competitive landscape and business strategies of Flock and Vidizmo.
“I think that this is, again, the reason I think it's such an important story to tell is that there's a lot of attention on Flock right now.”
Closing Thoughts and Subscription Promotion
37:45 to 41:45
Conclude with thoughts on the discussed issues and promote subscription benefits.
“but then they responded to several of my other inquiries about other stories last night, but not this one.”
Transcript
Automatic transcript. May contain errors.0:00Push your limits, train with precision, see the results. At Equinox, that's high performance loving. Everything you need to lock in and unlock your potential at Equinox. Start today at equinox.com. I think anybody who has touched this information is potentially a target. They really want to find out who is in the FBI specifically.
0:27Hello, and welcome to the 404 Media podcast, where we bring you unparalleled access to hidden worlds, both online and IRL. 404 Media is a journalist founded company and needs your support. To subscribe, go to 404media.co. As well as bonus content every single week, subscribers also get access to additional episodes where we respond to their best comments. gain access to that content at 404media.co. Also, do remember to subscribe to our YouTube channel where you can watch all of our episodes. Subscribe at youtube.com slash at 404media.co. I'm your host, Joseph, and with me are two of the other 404 Media co-founders.
1:10First being Sam Cole. Hey. And Jason Kebler. Hello, hello. Hello, Sam. Do you want to ask me about the FBI? Yeah, I'll ask you about the FBI. Our friend's the FBI. So this has been a story that's been like, I've seen it everywhere. People are sending me clips of people talking about it everywhere. It's been like the story for the past week. So let's break it down. The headline is, we hacked the FBI. Hackers say they have data on all FBI employees. And just to be clear, there are quotes around, we hacked the FBI. We did not hack the FBI. Hackers hacked the FBI. So how did you actually find out that hackers had hacked the FBI?
1:47Yeah, so I get a notification on my device. I pick it up. I see it's from Shiny Hunters, and they often ping me about new breaches, and they often ping a bunch of journalists as well. And the text message reads, quote, We hacked the FBI. We hold data on all FBI employees and applicants. It then goes on and describes the services they've hacked and that sort of thing. it says we've defaced an FBI website as well and it has a screenshot of that which is useful for verification purposes because it shows they actually did hack something as well. And then they say we have a bunch of data. Most importantly, they eventually give a sample of the data which I guess we can talk about in a minute as well.
2:37But the significance of the breach became pretty apparent very, very, very quickly. Like this hacking group has broken into an FBI service and they've stolen a bunch of data on what they say is all FBI employees. And this isn't just like some rando hacking group or a random person. This is someone that we know. So if listeners of the pod or readers of 4.4 and even readers of Motherboard probably will recognize the name Shiny Hunters. So who are Shiny Hunters? Who is the group? What else have they done that we might have heard of? Yeah, so Shiny Hunters, although they may disagree with this, so I think this is probably fair to say.
3:15They do come from the overarching, strange ecosystem of the comm that we've covered a bunch, which are these predominantly English-speaking hackers in North America, so Canada, America, some in the UK, and it looks like some spreading across Europe, and actually some in the Middle East now that we've seen a little bit as well. but they come somewhat from that ecosystem. And Shiny Hunters has been linked to a ton of different data breaches, not over the months now, over the years, right? They've kind of surfaced under different names. There's all of these different groups. Like back in the day, there was Lapsus, which got, I think, GTA data.
4:03And then there's Scattered Spider, which is a name given to a bunch of people who started the SimSwappers. Then you had Scatus Lapsus Hunters, which they called themselves. And then Shiny Hunters, or Shiny sometimes, has been around for a long, long time, operating around this sort of very nebulous community that's sort of very, very hard to pin down. But most recently, I believe, if I'm remembering correctly, again, with all these different names, but they were linked to the mass doxing of a bunch of ICE officials as well. And we covered that at the time because that was very interesting. It was sort of at the height of ICE's mass operations, sort of the mass raids.
4:48Usually what Shiny Hunters does is they will go into a company, they will hack it, they will extract data, they will then threaten to leak it to the public on their Tor hidden service we can go download it unless the victim pays, obviously. You know, pay hundreds of thousands or millions of dollars in Bitcoin or whatever. And if the company doesn't do that, they then release the data. And we'll talk about how this one's a little bit different later. But that's who Shiny Hunters is. They are, frankly, a prolific hacking group, collective, amalgamation. They have a lot of stuff going on, for sure.
5:25I crack up every time we talk about the names of these hacking groups. And I'm sure we've talked about it before on the podcast. but like you and Lorenzo Marteschi-BRI had done a blog that was like Shiny Hunters is like a Pokemon training gem or something. Well, they lean into that now. They have like a Pokemon like on their defense webpage. Oh, really? I didn't realize that. That's funny. So yeah, let's talk about what was actually in the data. So what are we talking about here when we talk about what was hacked? Yeah, so Shiny Hunters says they've hacked the FBI. I say that's very interesting.
5:59please could you provide a sizable sample of that because I need to verify this obviously first of all they send data on a particular FBI agent and their spouse and it includes their name address phone number the name of the spouse obviously immediately that's already pretty sensitive stuff like oh wow this is where this FBI agent apparently lives and that sort of thing. They also said that this agent was somebody that was part of the broader investigation against shiny hunters. So it shows that these criminals are thinking about, well, maybe we can find people in here who are investigating us.
6:38And in the past, we've seen people in this sort of community harass the FBI agents who target them. I've received and obtained a bunch of phone calls where hackers or criminals will phone up the specific FBI agent tracking them and harass them down the phone and like call them all sorts of slurs and names and all of that so there's that i needed more data though and then shiny hunters sent this spreadsheet this file of 5 000 fbi officials and their spouses and sometimes their parents as well i don't think we've actually mentioned that in the articles but yes it includes names multiple phone numbers addresses the spouse the parent, that sort of thing.
7:24And then that's sort of the personal side. Then there's the national security counterintelligence side where it includes like their job specifically. So I went through and then I found there were several people in there who were clearly linked to the remote operations unit, which is the really secretive part of the FBI that's basically his hacking team. These are the ones that are remotely breaking into phones, into computers. Back at Motherboard, I did some coverage on them where they've used classified hacking tools and normal criminal investigations. Like, the FBI does not talk about this hacking unit, but they were in the data as well.
8:04You then had Reuters go and dig through the data and found there were a bunch of FBI officials who investigate Russia and China in there. So this is obviously very interesting to foreign state actors where they can find out these are the specific agents who are tasked with investigating our country and maybe the criminal elements in us or maybe our governments as well. And then finally, there's also medical stuff in there. This wasn't in the sample I was given. Shiny hunters mentioned this early on in our conversation because they said medical data has also been taken. And I should say that the breach here sort of centers around an FBI job portal.
8:45And it includes applicants as well, but apparently a bunch, well, definitely a bunch of current employees as well. The BBC reported that the data also included like urine and blood tests from an FBI agent. And then Reuters added that the data also included like a psychiatric evaluation. So this is, I don't think you can get much more sensitive than this. both in the fact that it's the fucking FBI and also just how incredibly sensitive this data is, like on a personal level. Did they say at this point why they did it? It wasn't just for like shits and giggles. It was for a reason, right? Yeah. So...
9:27Sort of. Sort of. It's all very unclear and you do... Kind of all you have to go on are shiny hunters' statements and their actions, which I think we'll get to in a second, which is sort of what happened next. So the state of reason why they did it was that Shiny Hunter said the FBI previously published some sort of cybersecurity report, right? They do these publicly and they say, oh, you need to watch out for these hacking threats or whatever. In this case, the FBI said that shiny hunters, in their perception, often exaggerates the claims or inflates the claims of sensitive data that they have access to.
10:12The FBI said that shiny hunters will often send threatening text messages or phone calls to victims or their families to try to get them to pay. And they may sometimes engage in swatting. I'll say based on my coverage of the com that all tracks I don't know if Shiny Hunters specifically has ever done any of those things but that's what the FBI says Shiny Hunters they say they take issue with that they think this is inaccurate so they say the reason that they hacked the FBI and then they gave the FBI a one week countdown was for the agency to delete that report or to correct it It's basically like the most aggressive correction request sort of you've ever seen.
10:57And we get those occasionally, but people don't, you know, go this sort of far. We'll get to the exciting conclusion of the one-week countdown in a minute. But at this point, what was the FBI saying about the hack? At that point, the FBI was just saying, we're aware of these claims of a data breach on the FBI jobs portal. we're investigating it all of that sort of thing you know very standard within the first 24 hours or something statement saying okay okay we're looking into it we're looking into it yeah I'm sure they can't say too much at that point because obviously it's really it's a really up in the air dicey situation so yeah the next day you publish a story about the like you said the hacking unit was part of what was exposed which is a big deal because the hackers are hacking the hackers I would assume the hackers at the FBI are the best hackers in the country, I would think.
11:49I don't know about that, but they could probably do with some help. But it's still very sensitive. Yeah, for sure. And yeah, it's like hackers, hacking hackers is what we have at this point. So things are getting serious. And that was a story that you published the next day. And then a few days later, I guess this is where we get to the week is up or close to up. The shiny hunters people say that they actually won't publish anything. So the headline for that story was, FBI hackers say they won't publish massive trove of FBI employee data. So was it kind of like, we'll see if they do, or we were assuming that they might at that point, or we didn't know?
12:29What was the feeling until they go ahead and say, we're actually not going to? Was it like, oh my God, they might actually make this data public? What was going on in that interim? Yeah, so it's very hard to say. and in my opinion, I didn't think they were going to post it. I didn't think they were going to dump the data of all FBI employees. And now, of course, it looks very convenient when they've said they're not going to and I get to go, oh, I never thought they were going to in the first place. But I've been speaking to a few people about this and I'm pretty sure I genuinely believe this in that if I thought they were going to dump it, I would have pre-written an article.
13:09You know what I mean? Like sometimes we do that with like, something's coming, let's pre-write 500 words, and as soon as it happens, we can do it. I didn't do that because I didn't think they were going to dump it. And the reason is, is because there was this previous massive data breach at AT &T, right? Where hackers there stole the metadata records of all AT &T customers. One of the most insane data breaches of all time also came from this sort of overarching community. and I think there's connections between shiny hunters and the people who did that as well. In that case, there was a lot of boasting, a lot of threatening to release the AT &T data.
13:51But kind of what ultimately happened was that it didn't all come out. People got samples. Criminals still mined that data for information on specific FBI employees, just like they've been doing here with shiny hunters and the FBI data. But one person in particular released, I think, the records of Trump's contacts and family. And I think Rubio's wife was in there as well. So it was a very small, select sample they published. And sort of that was it. So based on that happening before, I thought the same was going to happen here. Well, yeah, they give samples to a few journalists. Maybe they post a little bit online.
14:29I didn't think they were ever going to post the full thing. Maybe that's just me being hopeful, because that would be absolutely insane if they published names and addresses of, was it how many people in the FBI? Like 38 ,000, something like that. Like that would be absolutely insane. Yeah, that would be pandemonium. And then they're also saying at this point, we never intended to do it anyway. It's like, oh, we were never going to do it. They said it was all a marketing campaign to protect our business and actively combat disinformation. So you've been dealing with these guys for a while. What do you make of that sort of statement from them?
15:06I do believe them when they said that we were never going to release it for the reasons I just said, right? But then you also have to take, basically, all we have is their word at this point where they're saying this is all a marketing campaign where we just wanted to correct the record that we are a threat or whatever. And that's sort of implied by their early demands for the FBI to correct this report, which to be clear, the FBI is not deleted or correct that report, and the FBI was never going to do that either. Shiny Hunter says the media ran with this and the public misunderstood it. Shiny Hunter says they never used the word deadline.
15:42I think they did use the word deadline in communications with maybe other reporters. I'm not sure if it was with me. But yes, it was. Well, on one side, it was a lot of, you know, flexing and boasting and that sort of thing. And I feel like some people may almost be like, deflated or kind of disappointed like oh so like nothing really happened in the end no no no no the fbi got hacked they stole this two to three terabytes of data absolutely insane breach they gave a bunch of that information the 5 000 officials to journalists that then somehow got out to you know john hammond who's a cyber security researcher and a youtuber i don't know how he got that information either from a journalist or cybersecurity researchers have their own way of getting this data sometimes and then made a video about that it's not like this breach happened and then no one got the data the data is floating out around there and i think anybody who has touched this information is like potentially a target by you know foreign intelligence agencies like russian or chinese who like they really want to find out who is in the fbi specifically if they don't already know.
16:59And the more this data proliferates and sort of spreads around, increases the risk that not only it might become public, but that it might just fall into, you know, some really bad guys' hands. So like, yes, they didn't dump it publicly. This is still like a massive national security and counterintelligence issue, in my opinion. Yeah, I mean, it's pretty bad just that the FBI can be hacked and that the hacking unit of the FBI can be hacked. That's not good aside from everything else it doesn't give me a lot of confidence in the security of i don't know the intelligence but i guess yeah the i hadn't even thought about the the aspect of like everybody who had who has this data now could be a target also it's easier to hack a relatively random person who got hold of the data or a journalist or whoever presumably then breaking into FBI servers.
17:54So now you have like a bunch of squishier targets potentially who have this information that other people may want. All right, should we leave that there? And I'll take us to the next story. After the break, we're going to talk about, well, a really significant scoop from Jason about facial recognition and Flock. We'll be right back after this.
18:20there's a point in the afternoon where my brain sometimes decides the workday is over even though the workday is very much not over and i don't always want another coffee especially if it's late because i know i regret it honestly if i've cold brew after like 11 a.m i'm up all night that's where ultra pouches have been really interesting for me here's my ultra pouches We have the tropical flavor. These are nootropic pouches designed for focus and energy. And the big difference is they have zero nicotine and zero caffeine. Ultra partnered with neuroscientists who formulate them with ingredients like Infinity PX, L-theanine, Alpha GPC, and vitamins B6 and B12.
19:01They're designed to give you smooth energy and focus that lasts around one to two hours. I've been using Ultra off and on for the last few months. And I really like them, honestly, when I wake up. And then again, before that afternoon crash, I feel like when I was just using coffee to get through the day, I was always really anxious and jittery. And what I like about Ultra is that I can keep it near my computer. And if I need a little pick-me-up, I can take it. But again, it's super smooth so that you don't feel jittery. It doesn't keep me up at night. It's a very calm, focused energy. There's five flavors, Cool Mint, Wintergreen, Tropical, Watermelon, Blue Raz.
19:38My favorite, again, is Tropical. So anyways, check out Ultra. It's the ultimate guilt-free pouch, delivering instant focus and mental clarity without nicotine or caffeine. New customers can use code 404media to get 15 % off at takeultra.com. That's takeultra.com for 15 % off with code 404media. After you purchase, they'll ask you where you heard about them. Please support our show and tell them that 404media sent you. This podcast is brought to you by NordLayer Browser. NordLayer Browser is an enterprise browser that secures online work without getting in the way of productivity. It gives IT teams full visibility and control over every web-based AI tool, SaaS application, and download across the organization while making sure employees can still enjoy a seamless browsing experience.
20:27Designed for today's distributed workforce, NordLayer Browser helps companies protect access to their internal resources and prevent accidental data exposure. Go to nordlayer.com slash browser slash 404podcast and use the coupon code 404podcast10 at checkout to get NordLayer browser at 10 % off today. That's nordlayer.com slash browser slash 404podcast and use the coupon code 404podcast10. All right, we are back. We have to take a quick break because Jason's mouse had died because of the amazing design of the Apple mouse. Get a battery. No comment. Let's talk about our stories. Yes. Okay. Serious business.
21:10The headline of this one is Surveillance Company Tells Cops It Wants to Add Facial Recognition to Flock Cameras. I mean, this is huge, Jason, because Flock very explicitly says it does not do facial recognition. It says it does not want to do facial recognition. at least at the moment, when we got sent data extracted from a Flock camera, I dug through that for evidence of any sort of facial recognition capabilities and basically found nothing. But the fact that Flock says it won't and it isn't doing that still leaves a gap in the market, right? That maybe another company could come in and essentially slap facial recognition onto Flock cameras or their data.
21:57That is what you have found here. Enter Vidizmo. Briefly, what is this company offering to do? We'll get into more on the specifics in a bit. Yeah, so I guess first, Flox CEO Garrett Langley has been like, we will not do facial recognition. Like here's a short clip of him saying this. We will continue to make commitments to the communities we serve. Commitments that may differ from other companies in this industry. We will not sell ALPR data to third parties. We will not add facial recognition to our devices. And we will not weaponize our drones. And so essentially, Vadismo saw this. And they're like, well, we'll do facial recognition.
22:37We do facial recognition already. We can connect our facial recognition systems to Flock OS, which is basically the back end that Flock uses. It's its real-time crime center. And they essentially propose to, I mean, they describe it as, quote, close the gap between what Flock can do and what police want Flock to be able to do, which is in some cases, facial recognition. You know, Flock already does object recognition, and they do some predictive type things. And they definitely like, lean heavily into AI. And I actually talked to the CEO of this Spadismo company, which we can talk about more in a second.
23:14But he was like, Flock is not handling this privacy conversation very well, which is a very funny thing for this company to say. But they're basically just like, this is the way that the world is going. They should lean into it. Of course, we want to offer facial recognition on this type of technology. And this follows a trend that we've seen over and over again, which is where companies will say, companies that do surveillance will say, we'll do X, Y, and Z, but we won't do C. We won't do facial recognition. We won't marry one type of data with another type of data. They will draw their own lines in the sand or their own red lines for where they think their product becomes too privacy invasive.
24:04And then what inevitably happens is other companies see that gap in the market and they say, oh, well, we'll do it. And they can then differentiate themselves in that way. Yeah, I think a really good example of that is Clearview AI, for example, where you had Facebook and Google who could do a bunch of facial recognition stuff, but they didn't want to release that product. You did have facial recognition for cops, but we built on sort of their own databases or a proprietary database from the company. clearview crossed the line in like well there's all of this you know social media and venmo profiles and all of this stuff out on the web we're going to scrape our masks and we're going to make that available to cops this feels like a similar thing um as in filling a gap but for flock and facial recognition so how did you first hear about this company it comes from an email Right.
25:00Yeah. So basically a deflock group in Johnson City, Tennessee, which sorry, it's from the wagon wheel song. It's like perhaps it's a very small, small city in Tennessee, but they have a deflock organization there sent me this trove of documents that they foiled and I was going through it. They actually told me to look into something else. And while I was going through it, I found this Vidismo email where they, a salesperson for Vidismo reached out to Johnson City and said, hey, you guys have really championed Flock. You seem to be really into Flock. We want to help you make Flock more powerful.
25:39So I think the actual email is pretty important. So they said, quote, Flock safety generates plate reads and clips continuously. Your investigators use that data on active cases, but that Flock data does not connect automatically to your axon evidence system. When a Flock hit becomes relevant to a case, a plate read, a clip, someone is pulling it manually from Flock and handling it separately to the Axon ecosystem. Vidismo Intelligence Hub closes that gap. It brings Flock safety data, Axon body-worn camera footage, and any other evidence source into one searchable platform. Investigators search across all of it simultaneously by face, vehicle, or object in seconds.
26:19So they're explicitly saying you can take this Flock footage or license plate photos, whatever, and you can search by face. And then if you go to the Vidismo website, they have technical documentation about how this all works. And crucially, they had a page talking about how they would integrate with Flock. They said, here's how we would do it. We would connect to the Flock OS backend, that sort of thing. And then they also have pages and pages about facial recognition and the things that they can do with facial recognition and how their facial recognition system works. And this is already a company that's offering facial recognition to cops for cameras that they already have.
27:01And so this is like an evolution of something that they are already doing. Yeah. Crucially, you said in the technical documentation, it describes how the company would ingest flock data. Just to clarify, what is the status of this? Is it like a pie in the sky plan? Or is it like, hey, they have some concrete steps to do this and they're pitching it, but they haven't quite implemented it. Like what's sort of the status? Yeah, so they don't have any contracts with police departments specifically to connect flock data at the moment. They are pitching it now. And I think that's notable that they're actually going to police departments and saying like, we can offer this.
27:43They told me that they don't have any contracts for this specific connection at the moment. So it's not happening. But there's basically two ways that it could work. And the first way could be done now where these Flock cameras are recording, cops can export that data and then they could just manually put it into Vidismo system or any other facial recognition system and run facial recognition on it. This is something that I've seen cops do, by the way, not with Flock, but with ring camera footage. There has been court records and documents I've seen over the years where it's like someone recorded something on a ring camera, that footage was sent to cops, and then they ran it through Clearview, for example.
Read the full transcript
28:28so that can be done now and that their ceo you know said like this is something that cops could do on their own kind of like without our help and then we would need to build this like automatic back end that would then pipe into like various real-time you know analysis software and things like that the other thing that he said was that uh like they just started kind of pitching this and that there's usually a three to six month lag time between a new sort of surveillance product being offered and then cops signing up, it being implemented, it being rolled out, all of that sort of thing. So again, this is not the only company that we've seen try to differentiate itself from Flock in some way, either by offering ALPR with live video, offering it with more privacy than Flock offers.
29:18In this case, Vidizmo is offering to do more than Flock can do or is willing to do. So I think that the reason this is important is because this is the type of thing that we've seen over and over again. It starts as an idea and then suddenly it's rolled out and cops are like, oh, well, now we have this. Yeah. And imagine that one police department does it and they manage to run facial recognition tech on their flock data. And I don't know, maybe they have 20 cameras in their town or something. And then another police department hears about that and then the company gets a testimonial and then that spreads.
29:53We've seen this happen over and over and over again with all sorts of technology. I can absolutely see this taking off if Flock doesn't intervene, and we'll sort of get to that in a minute. Just briefly on the technical stuff, like I've mentioned and you've mentioned Clearview AI a few times, the way that works is that they have made a massive database of faces by scraping the web. In this case, with Vidismo adding facial recognition capabilities to flock data, they, they say, are not providing faces, right? Like that would be up to the police. This is a little bit abstract, but how would sort of the database of faces work when it comes to this and flock cameras?
30:37Yeah, so basically the cops need to enroll the faces, which is pretty common for facial recognition systems like Clearview. One thing that it does is it searches the entire internet. What other facial recognition systems do is you upload a mugshot gallery or something. There's different databases of faces that cops can upload. And so they would be then adding faces to a watch list, more or less. And what Vidizmo says is like, okay, if you upload these faces and you tag a name to it, anytime that person is seen on any of the surveillance footage that you have access to, you're going to get an alert, there's going to be a box around their face with their name, and you can set it to automatically start recording that person.
31:22So you can basically automatically track that person anywhere that they go in front of any surveillance camera that cops control, which includes often, as we've seen, private surveillance cameras that the company, like the corner store or the hardware store or whatever, offers to send to police, which is increasingly common. You can then send a confidence score where if the algorithm thinks it's like 60 % likely, it's this person. You can set that sort of confidence interval of whether you want an alert or whether you want it to start recording. There's also the ability to like, if the system says that this is a marginal photo, like a blurry photo or something, you can override that, which I didn't include in the article.
32:08But basically, it's like they can click and be like, tell me anyway, which is quite problematic. And then also this company offers racial analysis, gender analysis, and age analysis, which are very, very, very controversial. The big no-nos. Yeah. I mean, for obvious reasons, but also to go over them, there's variance between races. Companies have spent a lot of time trying to make it more accurate on black and brown faces, which is like a double-edged sword there, but it's still not that accurate of a technology. This happened after there were several false arrests where the wrong person was arrested in Detroit, and I believe in New Orleans there was a case as well.
32:51They're basically letting the cops search by race. So you can go into the system and say, show me all of the Latino males age 30 to 50, and it will pull up footage of that. And it's just like, what could possibly go wrong? So many things could possibly go wrong there. It's such a mess. And it's very wild that that's being offered. Again, I asked the CEO, Nadim Khan, about this. And he was basically like, well, it's the same technology as object recognition technology. This is where things are going. So we just want to offer it. I mean, he was like, we have privacy and accountability stuff built into our system.
33:34which stuff yeah i mean he was basically like everything is recorded so like if cops are misusing it it can go back and they can go back and figure out what happened but like so does flock it's been abused time and time again so that that is not that's like a very small um small thing there right uh you mentioned that interview the ceo khan let's just play a little bit of that audio so people can hear sort of these comments for themselves. I guess facial recognition of Flock cameras is something that Flock itself said that it won't do. Like they said that they don't want to do it. Why is this something that's important for Vidismo to offer?
34:13Is it something that police have been asking for or that seems like it has investigative use? So first of all, you are focusing on facial recognition while If you look at the broader capability of the technology, the same technology is used to identify objects, items, physical things. And now it has a name, which is called physical AI. Physical AI means anything happening in the physical world translated into the virtual world. it is happening in schools colleges universities where all of these technologies are already implemented i think flock cameras is trying to get out of the way rather than trying to implement the technology right by by saying that they will not do the recognition but i think this is the world is going to be will have to be but with the right set of technologies where privacy and security are fully balanced.
35:14Jason, so you've touched on his comments a bit about the race stuff and sort of their privacy protections and just them believing that they're filling a gap in the market here. What did you sort of come away from that conversation with? What was your impression of this company and what they're trying to do? And what was the vibe you got? Yeah, I mean, it struck me as a small company that is trying to differentiate itself and doesn't care about the heat that will inevitably come from this if it's rolled out. I think that this is, again, the reason I think it's such an important story to tell is that there's a lot of attention on Flock right now.
35:59This company doesn't make ALPRs, but it makes the analysis software. And so it's like, they're willing to do something that Flock is not willing to do. There are cops who want to do this. Definitely, there are cops who want to do this. And it's like, the way that surveillance technology works is very often cops pilot or try these things out with very little oversight. And it can lead to really bad outcomes. I think that they struck me as a company that's doing a bit of a Wild West situation where it's like, well, we're just going to do it and see what happens. But at the same time, this company has been around for a few decades.
36:36They're a video analysis and database company. They're sort of new to surveillance, I would say. That's a pretty new product. It's like six months old. And so often the companies that are doing this sort of thing are brand new companies that pop up out of nowhere. Some of them don't know what they're doing. Some of them never end up offering a real product, even if they pitch it. Whereas here, this company has a long track record of making these video analysis products for governments. But they've only recently gone into law enforcement. And so it does seem like they have the technology, but I don't think...
37:18I didn't get the sense that they have the sophistication around how this could go wrong, the privacy impacts, the potential for misuse and what that could mean for them, etc. I've definitely got that sort of vibe from other smaller players like in the surveillance field or those who have just entered it. So finally, has Flock responded? No, and I've written obviously a lot about Flock. They usually do respond to me and they hadn't for a bit, but then they responded to several of my other inquiries about other stories last night, but not this one. I think... Are they just busy? I mean, they're busy.
38:00There's been reporting that they've offered employees layoffs and stuff because things are maybe not going so well for them. And so, I mean, that's possible. I think that they've changed their PR strategy several different times. So I don't know what's going on there at the moment. But I think that, I think Flock would say, like, we may be able to block this in some way. The thing I'd say to that, though, is that Flock has said over and over and over many, many times that police departments, the customer owns the data that their cameras record, like that is in the contracts. They've done that a lot of times to sort of like push accountability off to cops.
38:46And so if that's the case, it's not clear to me that they can actually stop this. They may be able to change their API in some way so that you can't export to this specific system. But I don't know how that jives with the customer owns the data and can do anything that they want with it. They can export it if they want to. I've been working on other stories about how cops do export this data regularly and quite automatically and often. And that often it's third party companies that are building the tubes, the piping that basically takes it from Flock and puts it into another system. And so I guess I'm just like, again, we don't, we kind of don't have the reporting into like exactly how Flux API works and export systems work and the code there and that sort of thing.
39:42So, I mean, it may be the case that Flux says we can stop this, but it would kind of change how they've treated data to this point. Yeah, or for Dismo, it would make some sort of hacky workaround and that sort of thing. But you're right in that Flux really, really stresses that the customer's control of the data and that sort of thing. And that makes sense because cities, really, and police departments, they want to feel like this is our collected data. It's not going off to Flock and they're not in control of it. And then there's sort of the double-edged sword of that. Yes, well, Flock doesn't want to hang off of facial recognition.
40:17But if you were truly saying that data belongs to the customer, well, they can do whatever they want with it, as you say, and that might include facial recognition. And I think if Flock turned around and said, well, no, actually you can't do facial recognition on this data. A lot of law enforcement customers will be pretty pissed off because they'll be like, this is our capability. This is our investigation. We should be able to do whatever we want with it. All right, really, really interesting piece and we'll obviously keep a very close eye on this company. If you are listening to the free version of this podcast, I'll now play it out.
40:51But if you are a paying 404 Media subscriber, we're going to talk about Meta's new AI agent, Muse, and how in some cases, it's actually just a human being making a phone call for that restaurant reservation of course it is. You can subscribe and gain access to that content at 404media.co. As a reminder, 404 Media is journalist-founded and supported by subscribers. If you do wish to subscribe to 404 Media and directly support our work, please go to 404media.co. You'll get unlimited access to our articles and an ad-free version of this podcast. you'll also get to listen to the subscribers only section where we talk about a bonus story each week this podcast is produced by alissa midcalf another way to support us is by leaving a five star rating and review for the podcast that stuff really really does help us out this has been 404 media we'll see you again next week
From the publisher
Don’t sleep on [@ultrapouches]. New customers get 15% Off with code 404MEDIA at takeultra.com! #UltraPouches #ad
We start this week with Joseph’s stories on the massive FBI hack. 404 Media broke this news and continued to cover the breach of data on all FBI employees and their spouses. After the break, Jason tells us how a surveillance company is telling cops it wants to add facial recognition tech to Flock camera data. In the subscribers-only section, Jason and Joseph chat all about Muse, Meta’s new agentic AI, and how humans are actually behind some of the tasks. Because of course.
‘We Hacked the FBI:’ Hackers Say They Have Data on All FBI Employees
FBI Hack Exposed FBI’s Own Hacking Unit
FBI Hackers Say They Won’t Publish Massive Trove of FBI Employee Data
Surveillance Company Tells Cops It Wants to Add Facial Recognition to Flock Cameras
Subscribe at 404media.co
Learn more about your ad choices. Visit megaphone.fm/adchoices
