How Police Are Watching You on Social Media
Documents from Chicago's Cook County Sheriff’s Office reveal the undercover techniques law enforcement uses to monitor—and manipulate—social media users.
In October, the ACLU released emails showing that a social media monitoring company called Geofeedia had tracked the accounts of Black Lives Matterprotesters for law enforcement clients. The revelations of social media spying made headlines and led Twitter, Facebook, and Instagram to cut off Geofeedia’s access to bulk user data (which in turn prompted the company to slash half its staff). Since then, two more social media monitoring companies, Snap Trends and Media Sonar, lost Twitter data access for similar surveillance activities.
Civil liberties advocates have celebrated these decisions, but new documents suggest police still have plenty of other tools to spy on social media users.
Jennifer Helsby, co-founder of the police accountability group Lucy Parsons Labs, provided CityLab with a slideshow prepared by a former employee of the Cook County Sheriff’s Office Intelligence Center that sheds some light on how police use social media. The presentation shows intelligence analysts how to mine location and content data from Twitter, Facebook, and Instagram—and advises them on setting up fake accounts and assembling dossiers on persons of interest.
One tip shows sites such as Statigram and Instamap, which can help law enforcement analyze photo trends or collect photos on individuals in targeted areas. This example points to images of individuals collected using Instamap near the Cook County Jail, which the Cook County Sheriff’s Office operates, as well as images of a child, a young woman, and families in Chicago.
Other slides reveal more advanced monitoring techniques. Geofeedia, the presentation states, can be used to geolocate users and conduct a “Radius and Polygram search” of an area for social media content. Echosec, a lesser known tool can monitor and geofence users, which allows police (and marketers) to track and collect users’ posts as soon as they are disseminated within a bounded area.
These tools rely on individuals’ public social media posts, but the slideshow also explains that police can use “catfishing”—creating fake accounts—to get non-public social media data, even though such accounts are not permitted on Facebook, Twitter, and Instagram.
While social media surveillance is often thought of as targeting certain locations or terms, such as hashtags, the Cook County Sheriff’s Office records suggest that intelligence analysts are also compiling information on persons of interest for longer term retention, not just for “situational awareness” at public events. Here’s a sample “Intelligence Information Report,” for example, to collect photos and other information.
The presentation doesn’t get into whether there are limits on who can be the target of these operations, or what legal safeguards they are ensured. One slide mentions terms such as “probable cause” and “search warrant” but there is no explanation if or how legal procedures affect the monitoring. Some of the slides suggest this police monitoring is not necessarily focused on dangerous criminal suspects. For example, the presentation links to an ABC news clip featuring a specialized LAPD unit “dedicated to tracking teen parties in real time by monitoring social media.” (The Cook County Sheriff’s Office declined CityLab’s requests for comment on its social media monitoring program.)
“As long as you type it, for police it becomes real.”
Joseph Giacolone, a retired NYPD Detective Sergeant and professor at John Jay College’s Law, Police Science and Criminal Justice Administration Department, says that while these undercover social media accounts may violate the terms of Facebook and Twitter, that doesn’t make them illegal. “It’s no different than running an undercover operation, or a buy and bust,” says Giacolone. “Requesting a friendship, as a policeman you have to be careful of that entrapment issue. But if you just put a half-naked picture of woman in there, you’re gonna get in. I mean how hard is it really? They’re gonna invite you right in.”
Nationwide, experts say there is very little clarity on how often undercover online operations are carried out. Surveys suggest such activities are often left up the discretion of police officers themselves. “Seventy percent of the detectives using this are self-taught and like half of the departments don’t even have a policy or procedure on how to use it,” says Giacolone, citing a 2014 LexisNexis survey on the use of these tools. “So cops are working without a net, so to speak, and you are going to see lots of challenges on these things.”
The same survey found while only 48 percent of departments polled had formal processes for using social media in investigations, 80 percent of law enforcement officers reported that they felt that “creating personas or profiles on social media outlets for use in law enforcement activities is ethical.”
One major concern among civil liberties advocates is that such methods are unfairly targeting the general public, not those who’ve already committed a crime. The LexisNexis survey also indicated that 40 percent of law enforcement officers had used social media monitoring to keep tabs on “special events” and 67 percent of respondents believed that social media monitoring “is a valuable process in anticipating crimes.”
ACLU of Northern California policy attorney Matt Cagle, who helped break the news on Geofeedia’s surveillance of Black Lives Matter, worries that covert accounts lack judicial oversight. ”This new world of surveillance products shouldn’t give law enforcement a blank check to create undercover accounts and collect information on law abiding people,” he says. “By using undercover accounts, they are potentially friending multiple people and getting much broader access than a warrant to Facebook for specific information would allow.”
So, who are really most likely to be targeted by social media snooping by law enforcement? Brendan McQuade, an assistant professor of sociology at SUNY-Cortland who studies law enforcement intelligence operations, is concerned that these methods will be used to crack down on political dissent. “It’s bad criminal tradecraft to broadcast your stuff on social media, so I would think this is more geared to political policing,” he says. McQuade points out that data already available to law enforcement, such as phone company records, allow police to grab suspects’ association and location information far more efficiently.
Geofeedia advertised its usefulness on this front during what it called “The Freddie Gray Riots” in Baltimore last year. In promotional materials aimed at police departments, the firm claims that its product, which was used by Baltimore County Police Department’s Criminal Intelligence Unit, helped police “to run social media photos through facial recognition technology to discover rioters with outstanding warrants and arrest them directly from the crowd.”
Giacolone says that social media surveillance can be useful for crime-fighting, but general only for low-level youths, not serious players. “You got two types of people out there: the young kids, the look-at-me generation that posts everything online, and then you have the older crowd that has learned to use social media to sell drugs through anonymous social media and its difficult to identify them. Most of these young kids are just posting, like, ‘Hey, everybody, I just robbed the store on the corner, look what I got.’”
“If you’re black or brown, your social media content comes with a cost—it’s a virtual prison pipeline.”
Even if these spying operations were only limited to suspects in such low-level criminal investigations, civil liberties advocates warn that users should be concerned about the ways in which their data is being retained and interpreted by law enforcement. Over the last few years, the NYPD has relied on millions of social media posts to justify “gang” raids across the city, targeting neighborhood youth crews whose online discussions sometimes relate to violence in their communities. Subsequent prosecutions for these raids have been controversial because social media chatter about shootings, drops in individuals’ volume of social media activity, and online photos of young men flashing hand signs have been questionably interpreted as evidence of gang membership and involvement in violent criminal conspiracies.
Matt Mitchell, a security researcher with the racial justice organizationCryptoHarlem, notes that the NYPD has carried out these operations by building intelligence dossiers on social media users over years, as the CCSO appears to be doing. ”The police are saying, ‘I’m going to follow you everywhere you go, write down every word you say, and look at every picture you take’, and now with these undercover accounts they are your friend hearing everything you say in confidence,” says Mitchell. “If you’re black or brown, your social media content comes with a cost—it’s a virtual prison pipeline.”
Often the NYPD’s social media surveillance gang operations collect and sift through social media content from teens and pre-teens over years, only to be used against them in court much later down the line. The Cook County Sheriff’s Office’s retention of social media data through intelligence reports could enable similar prosecutions. “What you say online is not always real. It’s not the same as something police pick up on a wiretap, but as long as you type it, for police it becomes real,” says Mitchell. “If you look hard enough, you’ll find something, no matter who you’re looking at. Take this post today, look at this thing they did five years ago, put it together, and you can draw any conclusion you want.”