Facial Recognition Technology Is Aiding ICE in Sanctuary Cities

Cambridge, Massachusetts, took another step last week toward banning the use of facial recognition software amid privacy and civil liberties concerns after Oakland and San Francisco, California, and Somerville, Massachusetts, passed their own bans. Ordinances are now pending in Berkeley, California, and state legislatures in Massachusetts and Michigan as privacy advocates push for a nationwide ban.

Federal agencies and increasingly, local law enforcement, are using facial recognition systems to identify or verify a person from a digital image or a video frame. The technology is becoming more prolific, as the Trump administration’s Transportation Security Administration moves ahead with its plan to use facial recognition technology to process and board passengers and to increase security at the nation’s airports.

But it remains unclear how the sensitive data collected by facial recognition software will be shared with other branches of the federal government, particularly Immigration and Customs Enforcement (ICE). Even in a sharply divided Congress, both Democrats and Republicans have expressed concern over the dangers posed by facial recognition technology and support for a federal ban. In particular, Democratic lawmakers have pointed to the technology’s tendency to misidentify people of color, as documented in research by the Massachusetts Institute of Technology’s Media Lab.

“Any time in the last three to four years that any data collection has come up, immigrants’ rights … have certainly been part of the argument,” says Brian Hofer, who chairs Oakland’s Privacy Advisory Commission and helped lead the push for the city’s ban. “Any data collected is going to be at risk when [ICE is] on a warpath, looking for anything they can do to arrest people. We’re definitely trying to minimize that exposure.”

Oakland strengthened its sanctuary city protections last year after the city council voted unanimously on a resolution to bar any branch of municipal government from cooperating with ICE, later formalizing the resolution into an ordinance. The city council also unanimously supported a separate ordinance that prohibits data broker companies such as Vigilant Solutions, acquired this year by Motorola Solutions, and Palantir Technologies from sharing personal information with ICE. Additionally, the city has already passed legislation drafted by the ACLU that forces police agencies to disclose their acquisition and use of surveillance technologies to local lawmakers and communities.

Elsewhere in California, concerns over data companies sharing sensitive information with ICE has sparked backlash. Last year, the Alameda City Council voted against a measure to spend up to $500,000 for 13 new fixed-location license plate readers, citing concerns over federal data sharing with ICE. Council members in Half Moon Bay, California, recently postponed their discussion of license plate readers while the state conducts an audit of the systems and their use by law enforcement.

“Any data collected is going to be at risk when ICE is on a warpath.”

Moreover, ordinances similar to the one in Oakland banning cities from awarding contracts to ICE, data broker companies and others who provide data mining services have spread across the Bay Area over the last year.

Hofer tells Truthout, however, that in Berkeley, the facial recognition ban faces a tougher fight than it did in San Francisco and Oakland, as council members there are already trying to delay and water down the ordinance with carve outs for certain consumer and employer applications. “I’m not confident that we have a majority [of the Berkeley City Council] that will go for a full ban,” Hofer says.

The fight for full bans on the technology may be an uphill battle for now, but activists say ordinances like these are essential because the threat posed by facial recognition — including within sanctuary cities and states — is unprecedented. For example, Vermont, a so-called sanctuary state, has recently come under fire after it was revealed last month that ICE used facial recognition technology to mine state driver’s license databases in three states that offer licenses to undocumented people.

According to The Washington Post, Utah and Vermont allowed ICE to analyze millions of motorists’ photos without their knowledge. Meanwhile, ICE agents subpoenaed Washington State’s Department of Licensing to conduct a facial recognition scan of license applicants’ photos. More than a dozen states allow undocumented people to drive with legal licenses or driving privilege cards, with some of those states already allowing the FBI to scan photos.

Data Sharing Violates Sanctuary Laws

The Post’s revelations about ICE’s mining of driver’s license databases are just the beginning. Unregulated data gathering and sharing is helping ICE monitor undocumented people in violation of sanctuary policies.

Many jurisdictions that call themselves sanctuaries do not specifically ban or even regulate information sharing. Rather, they adopt the sanctuary designation because they limit local authorities’ cooperation with ICE detainer orders to jail suspected immigrants beyond when they would otherwise be released. Other jurisdictions simply interpret the designation to mean that they won’t hold suspected undocumented immigrants as they await legal immigration proceedings.

Moreover, ICE contracts obtained by the ACLU show that data gathered by cities and states on residents within their jurisdictions finds its way into ICE’s hands daily. In addition to its access to a commercial database containing 5 billion records, ICE also has access to 1.5 billion data points shared by more than 80 local law enforcement agencies across more than a dozen states through its partnership with Vigilant Solutions. Two of those agencies are the local government of Merced, California, and an Orange County fusion center. Both are in violation of California state law, which protects license plate reader data and prohibits cooperation with immigration officials.

It’s not just license plate reader data that could inadvertently be making its way to ICE in a sanctuary state like California. Data including fingerprints, arrest records and mug shots routinely find its way into ICE agents’ hands. In fact, the Electronic Frontier Foundation (EFF) recently asked the California Department of Justice to improve its oversight of the California Law Enforcement Telecommunications System, which links multiple law enforcement databases, citing a potential for misuse of the data by ICE.

For other U.S. jurisdictions whose sanctuary ordinances aren’t as comprehensive as Oakland’s and which don’t specifically bar information sharing, undocumented immigrants remain at risk of local law enforcement agencies sharing personal data obtained through surveillance technology with Department of Homeland Security fusion centers, the Drug Enforcement Agency’s High-Intensity Drug Trafficking Areas program or Joint Terrorism Task Forces. In the “sanctuary city” of Austin, Texas, for instance, when police scan a suspect’s fingerprints, the data is automatically entered into a database shared with the FBI and ICE.

To the extent that ICE is using other forms of surveillance technology such as cell site simulators, privacy groups suspect the agency is using the devices in cases involving arrests that occur at very specific locations. The devices mimic a cellphone tower in order to sweep up a suspect’s cellphone data. Since the federal agencies often conceal their use of the devices through nondisclosure agreements, privacy groups are hoping to learn more about how ICE may be deploying them from a future legal case publicizing those details.

“There are legacy systems of data that ICE is relying on that don’t even require access to these technologies.”

But whether or not certain law enforcement agencies are using controversial surveillance technology is often beside the point, says EFF senior investigative researcher Dave Maass. “There are legacy systems of data that [ICE] is relying on that don’t even require access to these technologies,” he says. In other words, the ways in which agencies have been sharing data has long posed a threat absent comprehensive regulatory limits.

Still, he warns that the prevalence of license plate reader technology in particular poses a threat as the technology trickles down into rural areas and as the commercial market for license plate reader data grows, with private companies collecting increasing volumes of data to sell to other companies and lenders.

“If you know the license plate of somebody who is an undocumented person, you can add them to the system and get real-time updates about where they are,” Maass says. “You can also ask the computer system to predict where that person might be, based on their historical travel pattern.”

Targeting Tech Firms

Resistance to technologies like facial recognition extends beyond California and Washington, D.C. Grassroots immigrant rights groups are also building campaigns on this front, targeting private tech firms that are working with ICE to track and deport undocumented immigrants.

Mijente, an immigrant rights group, has spearheaded a campaign targeting the Palo Alto-based Palantir after it was revealed that ICE has used its FALCON mobile app to collect information on undocumented people, and that the app played a role in recent workplace raids. The app, which works in tandem with the company’s integrated case management system, allows ICE agents to search multiple law enforcement databases that contain data on immigrants’ family relationships and previous border crossings. Documents show that since 2013, Palantir has made more than $62 million in contracts from ICE for its FALCON app.

In May, The Intercept also revealed that the company provided software integral to a 2017 operation targeting unaccompanied children and their families. The revelations contradict the company’s prior public statements that it is not involved with ICE’s deportation division.

“This is the future of state control, and we have to have our voices involved in it.”

Palantir was one of three firms involved in the 2011 “Team Themis” scandal, in which it worked with Berico Technologies and HBGary Federal to gather information on journalist Glenn Greenwald and organizations, such as the Center for American Progress, Move On, Move to Amend, U.S. Chamber Watch, Brave New Films and the Ruckus Society, among others. The firms proposed hiring former intelligence veterans and creating false documents and false personas to discredit activists.

The uproar over Palantir’s role in helping ICE has also spread to Amazon, since Palantir uses Amazon’s cloud to run its software. Some Amazon employees have joined with immigrant rights activists to protest the company’s ties to Palantir. Moreover, Amazon has sparked controversy for selling its facial recognition software to federal agencies. Documents revealed last year that Amazon pitched its facial recognition system to ICE specifically as a method for targeting or identifying undocumented people.

“This is the future of policing,” says Jacinta González, who is senior campaign organizer with Mijente, explaining that the cooperation of tech companies extends beyond immigration policing, encompassing larger state systems of surveillance and control. “We see this trend of how the tech industry has been aiding these systems in growing and attacking our communities.”

González tells Truthout that Mijente is working with local organizers to incorporate information about Palantir’s software into “Know Your Rights” information and trainings, as well as working with students at several U.S. campuses organizing to interrupt Palantir’s recruiting programs at universities. The organization is also working with partnering groups to organize petitions asking the company to break their contract with ICE, and to organize direct actions and flyer campaigns outside Palantir offices.

Meanwhile, says González, it’s important to see Palantir’s cooperation with ICE as symptomatic of a larger trend: Technology is playing a growing role in the prison-industrial complex.

“This would be the biggest shift in power between the government and the governed.”

“Whether we’re talking about e-carceration and ankle monitors being used to expand correctional control in the United States, or whether it’s algorithms deciding who’s going to be free and who’s going to be incarcerated or locked up pre-trial, or whether it’s data brokers who are getting information on the whereabouts of immigrants; we see that this is the future of state control, and we have to have our voices involved in it,” she says.

The recent bans against the use of facial recognition technology and other controversial spy tech are one site of that resistance. As support for local bans grow, at least 11 cities and one county have passed model ACLU ordinances that force transparency in local police departments’ acquisition and use of secretive surveillance technologies. Nineteen cities and two states have introduced the legislation.

Meanwhile, advocates remind us of what’s at stake in this fight — a technology that is becoming increasingly ubiquitous and unavoidable.

“I really think this would be the biggest shift in power between the government and the governed; [facial recognition is] that radical of a technology,” says Hofer of the Oakland Privacy Advisory Commission.