Companies like Nextdoor have long attracted criticism for giving neighbors a forum for racist, fear-mongering posts. But one app goes further, encouraging users to document crime near their homes. On Friday’s episode of What Next: TBD, I spoke with Motherboard’s Joseph Cox about Citizen, which describes itself as “a personal safety network that empowers you to protect yourself and the people and places you care about.”
Critics say the app, formerly known as Vigilante, doesn’t just try to make people aware of what’s going on around them—it’s set up to make them afraid. What makes it more worrisome is that what Citizen is telling you may not always be true. Take what happened in Los Angeles in May: The company used its app and livestream to encourage users to hunt down someone suspected—wrongly—of setting a fire. At one point, Prince Mapp, the app’s head of community, told livestream viewers, “We need you to help us contact him and identify where he is. We need the scent of his clothing.”
This wasn’t just a TV performance. Cox got access to internal Slacks and other documents that showed how the company CEO, a software developer and entrepreneur named Andrew Frame, was thinking about the manhunt in real time. In the end, users did not find Citizen’s wanted man. The next day, the LAPD arrested and charged a different man for starting the 1,200-acre fire. Citizen apologized. But Cox says that won’t be the end of Citizen’s story.
This conversation has been edited and condensed for clarity.
Henry Grabar: On May 14, a brush fire broke out on the West side of Los Angeles. How does Citizen come into this story?
Joseph Cox: Citizen started doing what it normally does, which is inform its user base of ongoing events. But then information made its way into Citizen about the identity of a particular person that Citizen suspected of starting the fire. This was a person currently experiencing homelessness, and Citizen published this person’s name, their photo, and put a bounty on this person’s head. Any Citizen user that could contribute information that would lead to this person’s arrest would get $10,000, and then eventually $20,000 and $30,000 dollars by the end of the night.
And all this is happening on this live stream? In addition to the normal interface of the Citizen app, they have this almost TV channel that pops up for users?
Yeah, Citizen has this relatively new product called On Air. It’s basically a pseudo cable TV show where they have a host and they’ll be talking and discussing the incident. There’ll also be interviews with people who are at the scene. So, during this self-created manhunt, they’re broadcasting this. Not just in push alerts, but essentially as a reality TV show.
Can you read me some of the things that the CEO wrote in the company Slack while this was happening?
He says, “Find this fuck, let’s get this guy before midnight. He’s going down. Breaking news, this guy’s a devil, get him by midnight. We hate this guy, get him.” He explicitly says, “This will be good for our metrics. If we go and find an evil person every day, this is going to be better for our user base.” That was the killer quote for me, the explicit admission that this was at least in part to increase their user base.
It’s like if COPS were broadcast live and the people in the show weren’t actually cops.
Right, I should clarify that much of Citizen’s information does come from police scanners. They hire people in the U.S. and also potentially in Nepal and in Kenya, who are listening to police scanners and then just typing them up. So a lot of the information does come from the public police radios that you can access. But of course, that qualitatively becomes something else when you are broadcasting it to a huge chunk of Los Angeles and explicitly linking a bounty to it.
How do alerts make their way into the app? How do they make sure that what’s going into the app is verified? We’ve seen with social networks that are focused on neighborhood safety, how they can become a breeding ground for racist paranoia. What did Citizen do to avoid that and to make sure that doing something as delicate as corralling all this public safety information, they were protecting people who might get caught up in it?
Citizen does publish a lot of stuff, and sometimes it turns out to be false. Often this can happen with shootings. On the police scanner they could say, “Shots fired.” They turn up and that wasn’t the case.
Maybe it was a car backfiring or anything else. One report was of a woman was shot. A Citizen user turned up and there was no evidence that happened. They’re stuck in that they’re heavily relying on the police scanner audio, but they don’t necessarily verify. They’re just reporting what the police are saying. It’s intellectually lazy.
At least as local reporters, if you ran the crime blotter, you would at least follow up with the precinct and say, “What came of that? Did a woman really get shot last night on the corner?”
Exactly. They’re not doing much reporting at all. With some of the outsourcing to Cloud Factory, which has workers in Nepal and Kenya, these are people who are very far away from what is actually going on, trying to cover hyper-local incidents. This isn’t even local news. This is hyper-local news. Workers overseas potentially don’t have the necessary context to interpret the police scanners like that. There is also apparently some cases of including information that would ordinarily be cut out such as when there is a domestic violence situation. Citizen’s internal policies, as they were described to me, is that they don’t publish the specific address where that’s happening. They will say a street or corner. There’s some attempt of anonymization there, but sometimes those slip through the cracks as well.
What we saw in L.A. the other week, it strikes me, it looks more like the original iteration of the app, Vigilante. And some of your reporting indicates that while Vigilante may have been banned from the app store and changed its name, that the actual mission didn’t change that much. And they still saw the long-term goal as being about users fighting crime.
Yeah, absolutely. I don’t think the goal has necessarily changed. It has just been how they present it and their moves into COVID contact tracing bolster that. But when you have the CEO screaming in these Slack chats to capture somebody not even related to this incident at all—it’s highly concerning, right? I don’t like to exaggerate in my coverage because I want to be absolutely fair, but I one hundred percent agree with that, this is an app driven vigilante force.
Part of what you’ve found in Los Angeles is that this is a business model predicated on marketing private security.
This is how I originally got more interested in Citizen. A couple of weeks ago, a Twitter user posted a photo of a black tinted vehicle with the Citizen logo branded across it. This was Citizen reaching out into the physical world, they were no longer just on an app and transcribing information. They are testing an on demand, private security force. So, you would use the app. You would say, I need help, and a car would come up to you and then deal with whatever problem you had. It seems like this is a company that is desperate for monetization.
It sounds like Citizen is trying to circumvent the role of the actual police, perhaps even the justice system. They’re essentially putting out a digital wanted poster for a guy on their own accord on the one hand, and then on the other, they’re also trying to be the people you call when you feel like you might be in danger. They can show up and maybe provide a faster and a more efficient service than the police, or at least I imagine that’s the pitch to users.
In the internal documents we got, they mentioned that the Los Angeles Police Department, when told about this private security test from Citizen, allegedly called it a game changer. The documents then name who specifically said that and the LAPD hasn’t given any substantial answer to that. At least in the LAPD’s eyes, there is a lot of property crime and they simply can’t get their officers to the location in time. Citizen could potentially fill that hole. But of course, that is a conversation that shouldn’t be happening in secret.
It makes you wonder, too, what was the end game of the manhunt on Friday night for the man suspected of starting the fire? What was going to happen if someone after that $30,000 bounty found him?
One of the main takeaways from here is that we were lucky that nothing more substantial happened. Some man had a bounty put on his head and all these push notifications went out, but nobody—as far as we know—found him or harassed him or worse. So, I don’t know if Citizen’s is going to put another bounty on someone’s head. They have publicly apologized for this and indicated that they won’t, but the fact that they did this in the first place shows that this is a company that is willing to push the boundaries.
Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.