When does a smart city become an overly-surveilled city?
In this first episode of City Surveillance Watch, a limited podcast series from Smart Cities Dive, reporter Kate Kaye explores the inherent dichotomy of data-hungry technologies that, while promising to make cities safer and more efficient, can also be considered forms of surveillance tech.
Today, as decision makers evaluate use of technologies such as license plate readers, public safety cameras, smart sensors, and pandemic tech like body temperature scanners, they must balance the potential benefits of seemingly benign systems with very real risks and costs.
In this in-depth episode, listeners will hear from tech providers, city staff and law enforcement representatives, policy makers, civil liberties advocates and activists about how cities are thinking about these technologies. They’ll consider risks and unintended consequences of data-centric tech, and probe the gray areas that lie between a so-called smart city and one that’s overly-surveilled.
Sources featured in this episode by order of appearance:
- Wendy Hood, parking enforcement officer, City of Eugene, OR
- Brian Hofer, chairman, City of Oakland (CA) Privacy Advisory Commission
- Tyler Chandler, captain, Mt. Juliet, TN Police Department
- Lee Tien, legislative director, Electronic Frontier Foundation
- Robert Berman, president and CEO, Rekor Systems
- Ursula Price, executive director of the New Orleans Worker Center for Racial Justice, former director of New Orleans Independent Police Monitor
- Ginger Armbruster, chief privacy officer, City of Seattle
- Dierdre Mulligan, professor, UC Berkeley School of Information
- Jameson Spivack, policy associate, Center on Privacy & Technology at Georgetown Law
“It’s just telling me which location is picking up the hit, the location of the vehicle.”
Wendy Hood is a parking services officer for the city of Eugene, OR – home to University of Oregon and the Oregon Ducks – about 100 miles south of Portland.
During her day-to-day, she cruises through paid parking garages and street locations downtown checking for parking violations and issuing tickets when she finds them. Her lime green electric Chevy Bolt is part of a small fleet of cars that are tricked out with cameras connected to a license plate reader system made by Vigilant Solutions.
There’s a monitor perched on her dash that shows camera images of vehicles and license plates, along with data including dates, times and precise lat-long locations marking where vehicles are parked.
In today’s parking world, this system is cutting edge.
But the same technology from Vigilant Solutions – which is owned by Motorola Solutions – is used by law enforcement across the country. And sometimes its use can lead to unintended consequences.
My younger brother and I were going home for Thanksgiving, and rented a car and were about thirty minutes outside of Oakland, and a deputy sheriff’s car pulled behind us on the freeway, followed us for a little bit, flashed his lights, so I pulled over into this really well-lit shopping center.
Brian Hofer and his brother Jonathan were passing through Contra Costa County, CA that November night in 2018.
And finally one of the officers gets on his microphone and goes, “I’m sure you have some idea about what’s going on.” No - it’d be great if you’d talk to us and check our IDs, which they don’t. They point guns at us and direct us to walk out of the car. They handcuffed me, put me in the back of one car. They next directed my little brother to do the same and as he was backing up, this one officer just, I dunno, went crazy. Threw him on the ground, started screaming at him, pointed a gun at the back of his head, in kind of a executioner style while he was on his knees. And this is all in full view of me, you know, I’m just sittin’ in the car looking through the window watching this, ya know, it’s really traumatic when you’re in that scenario. They handcuff him, put him into a separate car and as we’re sitting there, I notice this Vigilant Solutions screen in the front of the car.
Hofer recognized the name on that monitor in the Sheriff deputy’s car: Vigilant Solutions. The company came up a lot in his legal and policy work chairing Oakland’s Privacy Advisory Commission.
Being a guy that writes privacy policies for license plate readers, I know who Vigilant is, but at that point, I didn’t know there was any connection, ya know, I just happened to notice the screen. Finally one of the deputies comes over and goes we got a license plate reader hit that this car was stolen. And I was like, that’s ridiculous, I just rented it a few days ago, obviously I didn’t steal it. They eventually do confirm that it was a valid rental. It’s still a little bit in dispute – it may have been stolen a few months beforehand, but regardless, it had been placed onto the hotlist for stolen vehicles in California, so when the license plate reader scanned that plate, it says, hey, this is an active investigation, there’s a stolen vehicle. We were just the unfortunate ones that that happened to.
Hofer and his brother were detained for about 45 minutes but never officially arrested. In the end, the brothers sued several of the entities involved. They won a settlement. Neither Vigilant Solutions nor Motorola responded to multiple requests to comment for this podcast.
Back in Wendy Hood’s little neon yellow electric car, this same exact tech, it feels revolutionary.
This is still pretty new to us, this system.
Does this feel like awesome innovation? Are you, like, the envy of your peers in the industry?
Oh yeah, oh yeah, definitely. This is definitely a huge leap forward more than anything.
Hood says the system is far more efficient and quicker than the old-school way – they used to mark the position of vehicles in paid parking areas with chalk.
City management says it’s a lot safer for enforcement officers, too – especially at night or in wet or icy conditions.
Whether it’s license plate readers for parking, drones to assess the state of city infrastructure or thermal imaging scanners that check for fever in the midst of the pandemic...
Right now municipal decision makers are being asked to evaluate emerging tech that promises to deliver a better understanding of city mobility, streamline services, reduce costs - and create convenience and safer environments for residents.
But, the juxtaposition of Hofer’s arrest story and Hood’s innovation story – demonstrates – there’s an inherent dichotomy in these tools.
As cities procure technologies that gather and analyze data in the hopes of being more efficient and safer, many argue they also enable new and more insidious forms of surveillance.
They worry about the slow creep of so-called smart or safe city connected tech use. They argue that – especially without explicit oversight, policies and safeguards, these technologies can and will be weaponized as tools of social control.
My name is Kate Kaye.
In this first episode of City Surveillance Watch, a limited podcast series from Smart Cities Dive, we’ll discuss trends emerging as municipalities employ data-hungry tech – we’ll talk about things like software mission creep and pandemic tech.
We’ll dig into the issues cities and residents grapple with.
We’ll hear from city staff, tech providers, law enforcement, civil liberties advocates and activists.
They’ll talk about the tradeoffs, risks and those unintended consequences of these technologies.
And we’ll analyze what surveillance tech actually is and probe the grey areas that lie between a so-called smart city and one that’s overly-surveilled.
And in future episodes, we’ll delve even deeper into how and why cities across the country from Detroit to New Orleans to Eugene, Oregon are using these technologies, why some oppose their use, and how cities can learn from those leading the way in crafting meaningful policy to avoid unintended consequences of surveillance tech.
When does a smart city become an overly surveilled city? We’ll explore that and more. This is City Surveillance Watch from Smart Cities Dive.
To many people, surveillance is a dirty word.
Headlines spotlight false arrests resulting from poorly-designed facial recognition.
Activists rail against over-policing of people from low-income and underrepresented communities, through video surveillance, smart streetlight cameras and gunshot detectors.
Legislators and civil liberties lawyers warn in dire terms: The more surveillance tech - like drones or cellphone detecting stingrays that cities procure, the more likely we’ll morph into a Xinjiang, China-like surveillance state.
But despite all the handwringing, there’s really no shortage of forces encouraging cities to use surveillance tech.
In Mt. Juliet, TN – a suburb on the eastern border of Nashville – Mt. Juliet Police Captain Tyler Chandler says his department plays an active role on social media when it comes to educating people about tech they’re using and promoting when it leads to arrests .
You know, it’s during the COVID times, we’re doing a lot of webinars, a lot of Facebook Lives.
The department has drone technology, body-worn cameras and sometimes uses video footage obtained through a partnership with Amazon’s Ring camera network.
They even branded their license plate reader program; they call it Guardian Shield. If a hit leads to an arrest, it goes up on Twitter.
Every successful apprehension is posted on our social media. We also have a running list.
Chandler says the few people who opposed the license plate readers at first came around once they realized the crime reduction benefits.
Ultimately with policing, the community is the police and the police are the community, so we all work together as one to keep a safe community, so it’s important that the community supports efforts that law enforcement is going forward with. We are a very transparent police department; we educate our community on everything that we do. You know, most people criticize things that they just don’t understand completely.
Law enforcement often gets what you might call a helping hand from starry-eyed local reporters who are quick to praise the use of shiny new technologies to nab alleged criminals.
[Sound: Local news montage]
Even influential think tanks like The Brookings Institute tout the benefits of, quote, “safe city innovation.”
In its 2017 report - Benefits and Best Practices of Safe City Innovation – Brookings wrote, “There are many opportunities for cities to build their economies and promote social inclusion through public safety innovation.”
Brookings pointed to the economic growth, job creation, crime reduction and boost in home values that can flow from city investment in public safety tech.
The institute even included a scorecard. Researchers assessed cities across the globe using a 120-point scale, rating them according to their degree of public safety innovation. The cities with the most high-tech public safety programs got the highest scores.
Singapore was the only place that achieved a perfect score of 120. Brookings praised the city-state’s video surveillance and mobile policing, enabled through public-private partnerships. It celebrated the Singapore Police for their use of digital technology and a program allowing citizens to send law enforcement counterterrorism information via text, photos or videos.
OK, so what I’m trying to get at here is that very influential forces are propelling cities to invest in surveillance tech or technologies that - even though they seem benign by some measures – have very real surveillance implications.
The smart cities movement, really, is, it’s international, from the Singapores, the Hong Kongs, the giant capitals.
Lee Tien is legislative director at the Electronic Frontier Foundation, a nonprofit that sometimes teams up with the American Civil Liberties Union to fight the use of technologies they believe infringe on privacy and free expression.
There’s a lot of interest in surveillance to make urban life more efficient, to keep track of...they have a lot of data. And it’s true – they could do a better job of understanding the data and managing the city. But they can also go overboard, weaponize, it can be a tool of social control. I’d say you can be in any country and see politics turn your technology into something that is not so great for a lot of people.
Smart and safe city tech - it’s big business. Tech research firm Omdia says the global safe city market will grow from $21.6 billion dollars in 2019 to $35.8 billion in 2024. Analysts there say investments in video surveillance systems, tech used to store and manage data and spending on broadband connectivity to link up sensors and city infrastructure is what’s driving most of that growth.
The global market for license plate recognition alone is projected to double from $916 million in 2020 to $1.8 billion in 2026.
It’s no wonder tech firm Motorola Solutions acquired the parent of license plate reader maker Vigilant Solutions last year for $445 million.
Vigilant also makes body-worn cameras for police and offers facial recognition tools that can match images of faces captured by its video surveillance cameras in near-real-time to people’s faces in a watchlist.
Vigilant competes for law enforcement and other municipal agency clients with companies like Rekor AI.
Rekor - they’re a public company - just launched a system incorporating its license plate reader tech with sensor-based data and vehicle recognition artificial intelligence. This is from a Rekor promotional video.
[Sound from Rekor video]
Police can use Rekor’s vehicle recognition to access data on vehicle make, model and color, along with license plate information to – as Rekor puts it “assist in the apprehension of criminals, and supercharge the recovery of AMBER and Silver alerts.”
Many of Rekor’s current customers are law enforcement agencies - including the Mt. Juliet, Tennessee police department. But the company’s President and CEO Robert Berman, he hopes those relationships swing open adjacent doors at other agencies inside city government.
For the most part, the product we just released, RekorOne, is a platform that we hope to capture multiple agencies within the same government. So let’s say in a state it could be used by law enforcement for public safety. It could be used by the department of transportation for everything from fleet management of vehicles to congestion pricing. It could be used by other agencies like DMV for identifying and keeping compliant uninsured motor vehicles, expired tags, lapsed registrations, and now since COVID we’ve got a number of customers that have talked about the idea of recognizing plates that are foreign to their state coming into their state.
So, this is where the distinctions between surveillance tech and other stuff cities use for things like making parking more convenient or helping the transportation department manage road wear-and-tear - they can get murky.
Tech firms that already have products not only add new features to the stuff that’s already installed on electric poles or street furniture, they partner with other companies to expand capabilities of existing city infrastructure.
Take ShotSpotter. The company makes systems that sense the sound of gunshots then relay detailed information to law enforcement about where and how many shots were heard. The company partners with GE and Verizon to incorporate its gunshot detection tech into their smart streetlight systems.
Companies like Rekor or GE are creating new markets inside the city governments they already work with, finding additional customers for the same products - new ways to use existing city hardware or new ways to analyze the same data.
Investors and analysts, they’d call these smart business moves.
Others might call it mission creep, a gradual spread from surveillance for police use into surveillance for all sorts of other city management purposes.
Here’s Brian Hofer again.
You know, they’re trying to integrate all these technologies to become this big super store of surveillance equipment. And so they’re offering all the bells and whistles. So it’s becoming really difficult to buy technology that doesn’t have all these other alarming features bundled with it.
The EFF’s Lee Tien, he calls it the surveillance industrial complex.
You know, my first introduction to this whole sort of area of smart cities was going to an event out at, like NYU where the head of their urban sort of policy center was celebrating this, what I considered a very dystopian vision of New York City completely instrumented where they, you know, could...reading every meter, had sensors on every corner.
Tien suggests that our attitudes toward high-tech public safety and security here in the US, they were changed drastically after 9-11. In particular, it was a watershed moment that laid the groundwork for extensive surveillance tech in places like New York City, says Tien.
The way New York City was happening was sort of organic from the fact that, you know, 9-11 happened, so from the terrorism standpoint, they were already being colonized by say, you know, Microsoft … and Microsoft collaborative and data mining software or other companies that were saying OK, we’re going to help the new NYPD and NYFD. And it’s all very – at one level, that’s great. At one level we want that to happen. At the other level we want to make sure that stuff doesn’t go overboard.
Let’s not forget: the Department of Homeland Security was created in the aftermath of 9-11. That agency oversees, coordinates and funds a network of fusion centers in states and cities across the country. These fusion centers, they’re digital monitoring hubs for law enforcement, public safety and threat-related data collection and analysis.
But, you know, if 9-11 conditioned people in the U.S. toward a more willing acceptance of government surveillance for the sake of safety and security… the entire evolution of technology over the last two decades has propelled it with computational rocket fuel.
Think about it – today, algorithmic tech is being sold to cities that is only possible because of recent exponential improvements in computational power. Once out-of-reach and expensive, today it is delivered as always-on, cloud-based subscription software.
And that means these days some of it can be pretty cheap.
Consider this as one example - In the four years the Sheriff’s office in Washington County, Oregon used Amazon’s cloud-based facial recognition system, the agency says it paid Amazon less than $2,000 in total.
Plus, companies selling tech to city agencies use affordability as a selling point.
Take Rekor – Though it won’t give details on its pricing for government clients, the company sells subscriptions to its vehicle recognition software for as little as five bucks a month to individuals and homeowners and $79 for businesses.
The company even gives away teaser licenses for some of its software to law enforcement. Here’s a bit of that Rekor promo video again.
[Sound from Rekor video]
All right, so, what are the risks of surveillance when it’s enabled through data-hungry tech?
Well, organizations like EFF, the ACLU, the Surveillance Technology Oversight Project and others, they oppose the use of technologies that they say could turn people into guinea pigs for a hyper-surveilled, hyper-connected society where governments cede decision-making to potentially discriminatory automated systems and people are essentially like walking data generators.
Civil liberties issues are often central to their opposition. Systems that can be used to keep tabs on people’s movements and activities – reflected in data points that can be stored, analyzed, shared and repurposed for all sorts of unintended uses – not only can create invasive privacy infringements. They can lead to suppression of the right to worship, protest, assemble, or simply move about in spaces freely without fear of perpetual observation from some personal data gobbling Big Brother.
And when rights are violated, it can lead to some real, practical costs to municipalities.
According to information provided by the County of Contra Costa, When Hofer and his brother sued the County and its Sheriff’s Department deputies involved in that license plate reader incident, they settled, paying the Hofers $49,500. The rental firm, Getaround, was dismissed from the suit and they declined to comment for this podcast. And again, neither Vigilant nor Motorola responded to multiple inquiries to comment about the lawsuit or other things mentioned in this podcast.
It turns out that license plate reader incident influenced Hofer’s work as an activist. In addition to his role chairing Oakland’s privacy commission, he’s also executive director of Secure Justice. They’re a nonprofit focused on surveillance reform and sanctuary immigrants’ rights.
So he got to thinking about the trust we put in technology, and the implications that has for our justice system as a whole.
Here’s Hofer again.
Where I really think it’s boosted or influenced my activism is that it’s in supporting really well one of our talking points about this, you know, scientific veneer that surveillance technology and computers always work and always give us accurate information and they’re not racist. So if a computer told me that, hey, here’s a criminal, it must be true.
We’re now all considered suspects in a perpetual lineup. We have the burden now of saying, no, we’re innocent. And that’s not the way our justice system is supposed to work.
The few municipal laws covering surveillance technology that do exist today are founded on principles of civil liberties, racial justice and equity.
Oakland’s surveillance ordinance passed in 2018. It stated – quote - “throughout history, surveillance efforts have been used to intimidate and oppress certain communities and groups more than others, including those that are defined by a common race, ethnicity, religion, national origin, income level, sexual orientation or political perspective.”
And tech-enhanced surveillance, it has its roots in slavery. Simone Browne, a professor of African and African Diaspora Studies at the University of Texas at Austin, writes about the history of slavery surveillance in her 2015 book Dark Matters: On the Surveillance of Blackness.
Lantern laws in 18th century New York City required Black and indigenous slaves to carry lanterns – the technology of the time - to illuminate where they walked at night. This way they could be watched and prevented from assembling to rise up.
Today, some suggest that what Browne calls “black luminosity” is manifested in modern ways, for instance through high intensity lights trained on areas in neighborhoods that are typically home to low-income or people of color.
And as we’ve seen in the way some biased artificial intelligence and predictive policing software has been designed, historical practices and data are used to inform future surveillance tech decisions, perpetuating disparate impact.
Ursula Price has fought an expanded video surveillance network in New Orleans for years. Until 2018 she ran the New Orleans Independent Police Monitor, the city’s independent police oversight office. Today she is executive director of the New Orleans Worker Center for Racial Justice. Price argues that in New Orleans, past policing has dictated the focus of today’s surveillance in that city.
They have used their history of policing as the justification for where these cameras should be placed. So, what that means to us is, those who are over-policed in the past will be over-policed forever because every decision today is being made on the basis of the discriminatory policing that happened yesterday.
[Ad break → Sign up for the Smart Cities Dive newsletter, here.]
So, how do cities define surveillance tech?
In most cases surveillance tech is defined as technology that enables observation, identification and analysis of people through data collection.
Seattle, home to tech giants Amazon and Microsoft, led the way among U.S. cities in 2017 when it established a surveillance ordinance.
Seattle says surveillance tech is technology that observes or analyzes “the movements, behavior, or actions of identifiable individuals in a manner that is reasonably likely to raise concerns about civil liberties, freedom of speech or association, racial equity or social justice."
Back in 2017 we had a list of hundreds of technologies that departments did a self audit, all of the departments in the city brought all that to us, and said, here’s what we think might possibly hit the definition.
Ginger Armbruster is chief privacy officer of the City of Seattle. Her office oversees the city’s privacy program and its lengthy impact reviews of surveillance tech. The city’s master list of surveillance technology casts a wide net. It includes 26 technologies used currently by the city’s transportation, lighting, fire and police departments.
Things like license plate readers for parking enforcement are on the list. So are closed-circuit TV cameras, sensor devices that use mobile phone data to track travel throughout the city, and infrared thermal imaging cameras installed on police helicopters to locate crime suspects.
But Seattle’s surveillance tech list could have been a much, much longer, says Armbruster.
We had to go through these hundreds and we had to cull. We got ‘em down at one point to 167 — I remember that number —and then we kind of went through that list. But unfortunately, we had about six weeks to do this, so we had in some cases a very brief description of what this thing did. And we had to figure out, in a conference room – ‘cause we couldn’t do a full review, right? Takes too long – does this hit the mark or not? And we got it down to I think 29 technologies that we felt pretty confident hit the mark, and went with that. That became our master list.
Since then, the city culled the list further to a final 26 individual technologies that are subject to an impact review process. They removed a drone and a state aircraft that, it turned out, are not used.
And then there was a facial recognition technology that police haven’t used and just said, you know what, we don’t trust facial recognition technology to be accurate, we don’t use this. It was a tool in the tool-chest and let’s just take it off the list.
Oakland is another national leader when it comes to establishing data privacy and surveillance tech policy… but Oakland’s definition – which has been adopted by cities including Portland, Oregon and less liberal San Diego - views surveillance tech more through a lens of data collection and use.
It says – and I’ll paraphrase here - surveillance technology means any software or electronic device used, designed, or intended to collect, retain, analyze, process, or share data specifically associated with any individual or group. And it includes data types such as audio and visual data, location data and biometric information. That’s like the stuff gathered by facial recognition tech or gait recognition technologies.
Deirdre Mulligan helped Oakland develop and pass its surveillance tech ordinance. Until recently she sat on Oakland’s privacy advisory commission – the one that Brian Hofer chairs.
I think people often think about, oh, surveillance technology – it’s going to be the IMSI catcher, you know, or the drone, or, you know, the police coming in with a new set of video surveillance cameras.
Mulligan is a professor at the school of information at UC Berkeley.
That first thing she mentioned – IMSI [IM-see] catchers – that stands for International Mobile Subscriber Identity Catchers – they’re often referred to as stingrays. They’re devices sometimes used by law enforcement that mimic actual cell phone towers. So, they essentially trick mobile phones in their vicinity into connecting to them and then revealing their unique ID and location.
But while some in city government may only think of things like stingrays and drones as surveillance tech, Mulligan suggests the category includes a variety of seemingly mundane technologies.
And in reality a lot of the information that ends up being used to surveil populations can be like much more banal.
It can be license plate readers, it can be all of these systems in our lives that are methodically tracking the movement of people throughout particularly the urban environment.
And so the implications, I think the implications for citizens turn less on where technology is being explicitly used in the context of policing. I think we understand that police are engaged in surveillance in a way that can sometimes be very protective, right? Often people hear the word surveillance and they think it’s just a negative connotation. But of course as a parent, right, you’re surveilling your children all the time. And we want the police to have their eyes open and be present, etcetera.
But I think it is important to understand that when we are building sensors into the roadways or we’re, you know, building a new parking, smart infrastructure or whatever, that that is all collecting a lot of data that’s gonna be tracking people as well and so the privacy implications of that and the questions about misuse and where the data’s going, you know, are equally worthy of city attention.
Data does not exist in a vacuum. Data and patterns of information representing people in urban environments can be stored over time. And today, software and cloud storage enables city staff and police to more readily visualize, analyze, integrate or share that information with other city agencies or across local, state and federal law enforcement networks – like through those fusion centers I mentioned earlier.
Here’s Mulligan again.
People often talk about it as a mosaic, right, and you can, you know, the mosaic it can be that it’s more data points. It can be that it’s repetition, patterns. There are many ways that having data from multiple sources or across multiple times can yield much more interesting information, right. So knowing where I was one day is interesting. But kind of knowing the patterns that I travel over the course of the week might tell you things that are much more interesting.
Like, am I always someplace on Sunday morning or Friday night or Saturday? And of course because we have a lot of background information, it’s not just knowing the location, it’s also having information about what the semantic meaning of that location is, like, oh, that’s a church, or you know, that’s a mental health clinic, or whatever it is, right? So, the availability of multiple data sources and the availability of data over multiple days or hours, whatever, all of it can really change the risk. It can raise the possibility for different kinds of harms because it can leak different kinds of information about me.
Oakland categorizes an array of technologies used by city government as surveillance tech - like video cameras and facial recognition, predictive policing software and automatic license plate readers such as the ones used by the parking department in Eugene.
The EFF’s list of street level surveillance includes electronic monitoring bracelets, gunshot detection sensors, and biometric data tracking tools like tattoo and iris recognition systems. Other lists also include tech that gathers biometric data such as gait recognition systems that detect people by the way they walk.
So, it’s not just tech hardware like license plate readers, cameras or stingrays that fall into this category. Many also consider algorithmic systems that analyze data like predictive policing technology, used to forecast who is likely to commit a crime or where crime might occur, to fall under the umbrella of surveillance tech.
But perhaps no form of surveillance technology has garnered more attention lately than facial recognition.
Law enforcement organizations and tech industry lobbying groups have pushed for city use of facial recognition tech. They often argue it can be of value in solving robberies, murders or locating missing persons.
But if anything, facial recognition has been subject to intense criticism, mainly because it has been proven that facial recognition algorithms have racial and gender bias problems.
A study published in December 2019 by the National Institute of Standards and Technology - that’s a federal government agency - found that facial recognition algorithms are less accurate when attempting to detect the faces of just about anyone who is not a white male.
Concerns that facial recognition use can result in disparate negative impacts for already vulnerable communities have led to several bans on the technology across the country in cities including San Francisco, Oakland and Boston.
But why does facial recognition get so much attention?
I think face recognition maybe is an issue that really resonates more with people because it’s their face.
Jameson Spivack is policy associate at Georgetown’s Communication, Culture and Technology department.
EDITOR’S NOTE: Spivack’s title is incorrectly stated in this episode. Spivack is the policy associate at the Center on Privacy & Technology at Georgetown Law.
There’s something about having your face being tracked or your face being identified or whatever that really, like it really makes people uncomfortable in a way that maybe an ALPR license plate data wouldn’t. Not to say that one’s worse or one’s better, but it really affects people in a different way.
Amid all those bans on face recognition, though, Portland, Oregon’s is the strictest yet.
Legislators there in September unanimously passed a groundbreaking ban. Unlike other U.S. municipalities that only prohibit use of the technology by law enforcement and other government agencies, Portland’s also covers use of the technology in privately-owned places accessible to the public like banks, retail stores, doctors offices.
Portland’s ban wasn’t theoretical either. It actually had a practical effect on privately-owned facial recognition that was being used in the city.
[Sound from facial recognition system - “Please look in camera for entry.”]
That’s the greeting from a facial recognition system that guarded entrances at three Jacksons Food Stores locations in Portland. The convenience store chain said the system deterred theft in those stores – two of which happen to be in historically Black parts of the city.
According to the new law, Jacksons had to disable the technology by January 2021.
The Portland law came at a moment of heightened tension for the city. Black Lives Matter protests had gone on for months. Portlanders demanded an end to police abuse against Blacks and other people of color and wanted reductions in police funding.
Federal and state police had descended on the city in the name of property protection and law and order. Some worried law enforcement could be using facial recognition to identify protestors.
Already facial recognition and drones had been used by law enforcement in Baltimore in 2015 to monitor protests after the death of Freddie Gray while in police custody.
And there were several reports of aerial surveillance by federal agencies over BLM protests across the country using drones and other aircraft.
So, despite the fact that the Portland Police Bureau had not used facial recognition, the timing was apropos for a vote to prohibit use of potentially discriminatory surveillance tech in the name of public safety or law and order.
Use of surveillance technologies during protected activities like protests or gathering has a stifling effect, says Deirdre Mulligan.
It’s really hard to hide your face, right? And the questions of, like, what does it mean for public faces [spaces] to kind of be conducive to the kinds of activities we expect to occur in them, right, for people to feel free to protest or to gather or whatever. And as this technology proliferates, I think it’s raising a lot of questions regardless of whether it’s in public or private hands.
So, you know, it seems obvious that some stuff like facial recognition fits squarely in the category of surveillance tech. But there are lots of grey areas.
One of those grey areas? The Mobility Data Specification. The controversial standard for supplying e-scooter tracking data, it was developed by Los Angeles and has been adopted by other cities. Some say it, too, is a form of surveillance.
Cities say they want e-scooter providers to use the mobility data standard in part so that they can get meaningful data to understand city travel and help improve mobility services.
But the ACLU and EFF are suing LA over the MDS. They argue it requires e-scooter firms to cough up a level of data granularity that is just not necessary, like unique e-scooter IDs and precise locations where trips start and end. And they worry cities will extend the requirement to ride-hailing, car-sharing services and driverless cars.
The EFF’s Lee Tien acknowledges cities can derive value from mobility data when it comes to achieving equity goals, but he and others question whether they need so many details to get there.
It can come propelled by some very, very good causes, shall we say, right? So, we know that folks who are working on organizing labor and trying to work on wage equity are really, really interested in getting more data from, say, the Ubers and Lyfts of the world in order to understand how much driving time are drivers actually being paid for, etcetera, etcetera. So there are reasons you might want that data in order to address some issues. On the other hand, how much data do you really need to accomplish the important social goals you want?
That is a question, that is a question unfortunately that many governments don’t ask and they don’t want to ask. But it is exemplary of the thought process that we think actually needs to happen. People have to think about these costs and right now many of the cities do not and do not want to.
The latest smart safe city surveillance tech trend: pandemic tech.
COVID-19 has created a whole new surveillance tech balancing act for cities to manage.
Municipal public health and emergency services staff have been getting pitched on mobile contact tracing and triage apps, new data sources and other touchless pandemic tech – some developed recently by tech firms they might already have relationships with.
[Sound from Avigilon’s no face mask detection technology video]
That’s from a promotional video from Motorola-owned Avigilon. The company is selling governments and businesses new tech to monitor for face mask compliance. The AI-based “no face mask detection tool” alerts monitors in real time if cameras detect someone without a mask, and it uses analytics to identify times and locations where violations occur most frequently.
And then there are those increasingly popular cameras and scanners that use thermal imaging to detect elevated body temperature.
Some consider these thermal body temp scanners to be simply smart public health protection. They’ve been used in transit hubs and airports across the globe in response to other deadly virus epidemics.
But some question their efficacy and argue thermal temperature cameras are just the latest form of security theater.
The EFF talks about thermal temperature cameras as surveillance cameras. The group argues, “Spending money to acquire and install infrastructure like so-called ‘fever detection’ cameras increases the likelihood that the hardware will long outlive its usefulness during this public health crisis.” And they warned it could open the door to facial recognition.
It already has actually. Some of these thermal temperature readers do employ facial recognition. In fact, right now, Seattle city buildings are outfitted with thermal imaging body temperature tech that actually uses a form of face recognition.
Seattle does not have a facial recognition ban. But what about its surveillance tech policy? Ginger Armbruster explained why, at least according to the way it is implemented by the city, Seattle’s thermal temperature scanning software is not considered surveillance tech there.
We do have technology in place in all of our buildings. It’s part of the return-to-work program where we want to take temperatures before people come into the building.
The City of Seattle has put – and let me explain the technology. It’s on an app, on a phone, mounted on a pole and it does a scan of the forehead from a distance so that nobody has to touch you or be involved to do this, and then enter the building. It does facial recognition, but – pause for a moment — it does facial recognition to say you have a face and that’s where your forehead is ‘cause it’s trying to triangulate on that part of your face.
If it did do tracking, suddenly we’ve got facial recognition a concern, we’ve got who are you and your temperature. What we’ve done to make this privacy protecting and not surveillance is we do not have any of that tracing, tracking or data collection associated with this on the application on the phone as it is in the City of Seattle.
So, we’ve done a detailed review of that technology before it was implemented. They are starting to put them into different buildings throughout the city. We have the policies in place that say this is how you will configure these devices, and even if it can do all this you’re not going to put all of that in place. Sign on the bottom line and that’s how we are implementing this.
Privacy advisors in other cities aren’t so sure. Brian Hofer says city staff in Oakland wanted to use thermal temperature scanners in city buildings there like libraries and community buildings, but the proposed technology had a facial recognition feature. Oakland does have a facial recognition ban, but that’s not the only reason Hofer was against the proposed temperature scanning technology.
And so I’m going through all these materials, and I’m like, we have a ban on facial recognition in Oakland, you can’t acquire this technology. It says right here that they’re using facial recognition - because they wanted to make it easier if you were like a repeat visitor to city hall, like you worked there, that you could scan in and it would recognize you and maybe give you the green light faster. And it would retain data, it had Wi-Fi, it would be sending data to the cloud, the vendor was also in the storage chain. I was like, no, we can’t do this.
Secondly, your stated purpose was forehead temperature checker. Go get that. If that’s what you wanted, go get that. You’re getting all these other features that then become a civil liberties concern.
There’s no doubt about it. From license plate readers to thermal temperature cameras, these technologies present a host of thorny issues for already-overworked and under-resourced city agencies to navigate. And the pressure on them is coming from all sides.
Can municipal staff and legislators balance the potential benefits of technology with the very real risks of excessive or invasive surveillance?
Here’s Armbruster again.
It’s always a balancing act in smart city work or any data collection, trying to provide services between, what should we be collecting and how do we protect privacy, and so it’s a real balancing act. And that’s where I am, I’m at that balancing pivot mark. I’ve got departments over here, I’ve got folks with concerns about privacy over here, civil liberties groups, and we’re trying to find the balance. And so, all of this feels like it’s on kind of the leading, bleeding edge of discussions about how we provide advanced services, take advantage of technology but not let technology take advantage of us.
Next time on City Surveillance Watch
[Sound from City Surveillance Watch Episode Two]
Join me, Kate Kaye, as I delve even deeper into stories about surveillance tech: what’s happening inside gated communities like an HOA in Tennessee, stories about how and why police in Mt. Juliet, Tennessee and Kansas City use tech like license plate readers and data from Amazon Ring home security cameras. About surveillance and activist opposition in New Orleans and Detroit. About public-private surveillance tech partnerships. Parking enforcement innovation with surveillance implications in Eugene, and more.
Until next time