Issue 15 / Beacons

December 25, 2021
Abstract black and white interstitial image for The Oversight Bloc

The Oversight Bloc

Lilly Irani, Khalid Alexander

How community organizers and tech workers joined forces to rein in San Diego’s surveillance state.

In July 2019, one of us, Khalid Alexander, received a tip from a fellow San Diego community organizer. “You should be paying attention to the city’s new streetlights.” The message continued, “Apparently, they have cameras attached to them.” Alexander lived in one of the many predominantly Black and brown neighborhoods in San Diego that was under constant police surveillance, including by “gang suppression units” that watch, harass, and document residents. He feared that streetlights with cameras on them could supercharge these efforts.

Two weeks later, Alexander showed up at a public library for a forum about the streetlights program (which the city named the Smart Streetlights Program). The only other people at the meeting were the presenters: a police captain, a city staffer, and an executive from General Electric (GE), the company that produced the new streetlights. Their presentation began with an infomercial for the technology, a city-wide network of thousands of LED streetlights mounted with cameras that recorded video around-the-clock. The footage was uploaded to the cloud, where city agencies could use software to count cars, pedestrians, and who knows what else. According to the police captain, the smart streetlights were already being used to solve crimes.

Alexander left the presentation shocked and concerned. Five years earlier, police had rounded up thirty-three Black and brown men from San Diego and charged them with fifty years to life in prison for gang crimes simply because they were included in a police surveillance database called CalGang. The Smart Streetlights would make this kind of state violence more likely—other networked surveillance programs in cities such as Chicago and Los Angeles had led to similar raids—and yet none of the city’s larger civil rights organizations seemed to be visibly fighting the project. The community was going to have to organize resistance to the streetlights itself.

Alexander’s first call was to the Tech Workers Coalition (TWC). Much of the technology incorporated into the new streetlights was foreign to him, and he wanted to better understand how it might be used. He also reached out to activists from communities he knew were likely to be targeted by the surveillance—homeless advocates, immigration rights organizers, police abolitionists—and invited them to the next streetlights community forum. At that meeting, held a month later, Alexander and his fellow activists managed to pack the house.

This was the beginning of a two-year grassroots campaign to rein in the Smart Streetlights Program. By connecting surveillance to a wider set of issues than policing alone, the organizers were able to create a coalition wide enough to win. A lynchpin of their success was not only building a traditional coalition of community organizations from around the city, but also bringing in new allies: workers from the companies and research institutions building these technologies. The methods might be a model for other struggles against surveillance and carceral technologies in cities around the country.

The New Bacon

The infomercial Alexander was shown at the first community forum boasted that the data produced by the Smart Streetlights were “the new bacon”—they went with anything and could serve almost any purpose that agencies with access to the data could imagine. The other one of us, Lilly Irani, an organizer with TWC, listened to the presentation at the second community forum and realized that surveillance was only one part of what the streetlights could be used for.

Community members at the forum immediately raised concerns about the technology’s ability to further criminalize San Diego residents. Black and brown-led organizations worried about heightened racial profiling using video streams. Refugee advocates worried that the streetlights could intensify the criminalization of Muslims by using software to analyze behavior near mosques. Homeless rights groups predicted the city would use the streetlights to more quickly find encampments to sweep. Border activists worried the streetlights could help track and deport people by being integrated with systems such as those developed for the Department of Homeland Security by Palantir.

A police officer at the forum assured the crowd that the cameras did not record private property, but a computer engineer from TWC was able to force the officer to clarify that the cameras did record private property, which was then scrubbed from the data by a software program. What the city called the “sensors” on the streetlights also included microphones that could record people’s voices without their knowledge. The officer said the microphones were currently disabled, but admitted that they might be used for gunshot detection in the future. When a community member asked how long the streetlights were going to collect information, a city staffer replied that it was “undetermined”; the officer tried to reassure the crowd that “the answer is probably never.”

Activists recognized that what the city and law enforcement said it was doing (and planning to do) with the program was beside the point. San Diego has long been an innovator in police surveillance networks. The county’s Automated Regional Justice Information System, which shares surveillance data between more than sixty-five law enforcement agencies, has been in place since the 1980s. More recently, the county had provided facial recognition devices to law enforcement agencies across the southwestern border region until the state passed a moratorium on such devices and the Electronic Freedom Foundation sued to have the moratorium enforced. For years, the San Diego Police Department shared data from its automated license plate readers with Border Patrol to help the federal agency track, detain, and deport migrants, until the practice was exposed by journalists. When it came to the streetlights, the hard part was already done: more than three thousand units were installed across the city. Expanding what that hardware could help officials do would only be a few software updates away. As the officer admitted, the sensors weren’t even covered by civil codes the way similar technologies, such as automatic license plate readers, were.

And it wasn’t only city and law enforcement agencies that might deploy the technology. All of the data generated by the streetlights would be made publicly available, albeit in an ostensibly anonymized form. A city staffer at the forum explained that this would allow “civic entrepreneurs” to use the data to build businesses to improve the city, including an app to help people find parking downtown. Irani knew this was code for gentrification—the city would give or even sell data harvested from the public, without its consent, to tech companies that would use the data to reshape the city in ways that favored wealthier white communities and pushed Black, brown, and poor people out of San Diego. Irani and Alexander later discovered through a public records request that the city had used Federal Community Development Block Grants, meant for projects that benefit low-income neighborhoods, to fund the streetlights.

This range of concerns about how the technology might be used in the future became the basis for the grassroots campaign. Within a month, thirty community and activist groups had joined together in what became the Transparent and Responsible Use of Surveillance Technology San Diego (TRUST SD) Coalition. It grew quickly, in part, because people recognized that the Smart Streetlights surveillance infrastructure spanned the city and could bring anyone into the law enforcement dragnet. The campaign also relied on the working relationships that Alexander had built over years of organizing with different communities.

Tech workers, too, were an essential part of the coalition. They joined because they did not want to build tools of state violence and oppression. In many public contests over technology, experts from industry and the state dominate conversations by wielding technical knowledge that communities cannot counter. In this coalition, however, the technical expertise of tech workers allowed organizers to counter official claims about what the Smart Streetlights technology could or couldn’t do, and helped them understand how the surveillance applications they feared could in fact be built on top of the existing network. The alliance with tech workers neutralized the city’s advantage.

Organizing the Lab Rats

The coalition decided to focus on three goals. Articulating these was crucial to keeping the various political and ideological factions within the coalition from splintering over other issues. First, it called for an immediate moratorium on streetlight acquisition, installation, and operation. This would end the immediate threats that the streetlights posed while organizers lobbied for more radical changes to public policy.

Those changes to public policy were the coalition’s second goal. Organizers sought public participation in the creation of legally enforceable policies over all surveillance technologies used by the city, not just the streetlights. A wide range of technologies beyond the streetlights—known and unknown—made up the surveillance dragnet of San Diego. Worse, surveillance tech could come to the city through donations to the Police Officer Association or through free trials like those offered by facial recognition company Clearview AI. This meant that the City Council did not always have to approve the technologies and the public had no way to see what was in use, unless someone peppered city departments with scattershot public records requests. Council members had not even discussed the Smart Streetlights Program publicly before they signed off on it; that was two years before communities realized a mass surveillance technology was even on the table. Communities needed a legal infrastructure that would alert them to surveillance technologies before they were approved. Rather than playing whack-a-mole to stop individual technologies, the coalition sought transparency, oversight, and City Council authority over all city surveillance technology, existing and future.

Finally, the coalition demanded public records showing how the streetlights had already been used and accessed. As organizers studied these records, they discovered that the city had considered monetizing the data and had even explored providing a livestream to the police. These discoveries informed the coalition’s activism and provided fodder for breaking news stories that added momentum to the campaign.

Because the mayor had championed the program, organizers would need a veto-proof majority of council members to support the moratorium and the policy changes we were fighting for. So the campaign worked to build public pressure on the city council through press conferences, newspaper opinion pieces, public events, and direct lobbying. Organizers broadened their discussion of the issue beyond criminalization to larger concerns about the dangerous power of big tech to engage liberal and conservative elected officials in different ways. In one opinion piece, they argued that San Diego residents had been turned into “lab rats for innovation.”

The Mayor’s Office and city officials tried to dismiss the campaign as a small handful of activists with hidden agendas. They also attempted to frame it as ignorant of the technical dimensions of the streetlight program. In November 2019, a tech worker who had experience working on artificial intelligence contracts read the contract the city had signed with GE, and discovered that the city had signed over to the company ownership of the data produced by the streetlights. The data were stored on GE’s servers, and the city merely accessed them through a subscription, which meant the city lacked final say over what the company could do with the data. An anonymous city staffer responded by calling the coalition’s findings “insane lies,” but ultimately the city could not undermine the credibility of tech workers who had built or worked closely on smart cities hardware, artificial intelligence models, and similar technologies.

Over the following months, members of the coalition lobbied city council members and found them increasingly at odds with the mayor and the city attorney over the scope and legal framework of the Streetlights Program. In order to channel that frustration into a tactical win, organizers followed a two-pronged approach. First, they expanded their attempts to put public pressure on these elected officials. They held town halls, screenings of the film The Feeling of Being Watched—about government surveillance of Muslim communities in Chicago—and workshops at which we taught community members and activists about the technology and brought them into the organizing efforts. They also did direct outreach to community leaders and journalists.

The second prong was to write a “surveillance technology transparency and oversight” ordinance. The coalition would hand the draft legislation over to a champion on the city council, who could then present it as a “common sense” solution they could claim as a legislative victory. Organizers adapted their ordinance from ones already in place in Oakland and Seattle. It required city departments to create use policies and impact reports to gain Council approval for any new technology with surveillance capabilities. It emphasized oversight over all such technologies, not just those used by the police. The LED streetlights, after all, had been acquired in the name of energy savings and innovation, obscuring the technology’s connections to law enforcement. The ordinance also created a Privacy Advisory Board to support the City Council with recommendations on specific technology acquisition proposals. Board seats were reserved for representatives from “equity-focused organizations” serving communities impacted by surveillance, as well as information technology and civil liberties experts; anyone with financial ties to companies selling surveillance technologies was disqualified. 

Organizers wrote the ordinance to appeal to a wide range of constituencies. Fiscal conservatives liked it because they wanted to reduce government spending. (The streetlight program, budgeted at $30 million over a decade, was already seeing cost overruns.) Liberals who believed in deliberative process appreciated that the ordinance created an independent body to advise the city council and included civil rights oversight. More radical organizing communities recognized that they needed the ordinance in order to find out about new technologies if they were to have any chance of organizing against them.

The coalition was lucky to find a champion on the Council who was willing to take the ordinance through committees and do the behind-the-scenes work to get it passed. The Councilmember, formerly a civil rights attorney at the ACLU, also had close relationships with people in the coalition. Meanwhile, organizers kept pressure on the city to shut down the streetlights and adopt the ordinance by organizing people to show up at council meetings and the mayor’s office, and led telephone and email campaigns ahead of crucial votes.

The coalition also seized upon Covid-related budget cuts to defund the streetlights. Mid-pandemic, Irani assembled a dystopian hackathon with coalition-aligned students to prototype creepy examples of what the streetlight technology could already do. Irani thought the demos would help persuade the Council of potential harms, but the most important outcome turned out to be the discovery that the streetlights didn’t deliver the promised data for city planning. The coalition alerted an investigative journalist, who broke the story in April 2020. By May, the city was proposing Covid budget cuts. The mayor had put library hours on the chopping block but saved the streetlights. The coalition worked with the progressive Community Budget Alliance to mobilize residents to email and call council members and demand they defund the streetlights. San Diego Climate Action also joined the effort, since the lights, operated under the Sustainability Department budget, had redirected funds to a broken, greenwashing surveillance system. In the end, the Council refused to fund the system. 

Tools for Struggle

In May 2020, the mayor made a last ditch attempt to hand over control of the surveillance system to the police and fund it using an obscure budget pool that City Council didn’t control. The coalition was able to show through old legal memos that this funding strategy was likely illegal and held a press conference to make the point. The mayor fought organizers for months, but finally, in September, he surprised everyone by announcing that the Smart Streetlight sensors and networks would be turned off until an ordinance was created to oversee the surveillance technologies run by the city. 

In November 2020, a year after the campaign began, the Council unanimously approved the oversight ordinance. The ordinance makes visible technologies that usually operate out of public view, mounted on light poles and cop cars, or running in the circuits and servers of hardware, software, and data brokerages. It slows down technology acquisition and gives communities time to learn and organize resistance. It puts community members with negative experiences of surveillance in a role where they can build knowledge about technologies and educate others. Though some argue that such ordinances create legitimacy for surveillance technologies, they also create a mechanism for people to organize refusal where there currently is none. 

But ordinances like these are not a panacea. They are tools for struggle and refusal, but do not guarantee resistance to surveillance. Without vigilant organizing, including alliances with technologists and elected officials, even community advisory boards may rubber stamp policies and legitimize surveillance technologies. This struggle also shows how cities do not control the technology of companies they contract with. As the coalition in San Diego worked to get the ordinance passed, it put the fear in city council members by explaining how the NYPD lost control of its data to Palantir. Then, the same thing happened to San Diego. With defunding, the city lost access to the streetlights’ surveillance feed. But the cameras continue to record. GE sold off the streetlight network to another company, which sold it to a Florida-based firm called Ubicquia. Ubicquia refuses to stop recording even though the city can no longer access the data. But even this did not stop the police from removing the camera and harddrive in one case to access and share the video. 

Paradoxically, it was the process of organizing for the ordinance that strengthened the coalition’s political capacity to challenge emerging surveillance technologies. By political capacity, we mean relationships among community members who trust one another, can teach each other, and can work together; we mean the time that people can spend researching, calling into council, occupying the mayor’s office, strategizing, and running educational forums. More people means more time spent doing these things, and more relationships with people who will get involved in the movement. 

In advocating for the ordinance, Irani and Alexander talked to a wide range of people, from anti-racism activists to former soldiers. Each person had different reasons to fear surveillance. People who showed up out of fear of big tech or privacy violations also learned about the criminalization of marginalized San Diegans. Anti-criminalization activists learned a lot about technology and its potential role in further entrenching the carceral state. Many of the people engaged by the coalition took their first steps from awareness to action. But the coalition’s work to resist and refuse mass surveillance doesn’t end with the ordinance. It begins in earnest once the ordinance is in place.

This piece appears in Logic's issue 15, "Beacons". To order the issue, head on over to our store. To receive future issues, subscribe.