Certain cities and communities are watched much more closely than others
Follow along for more content at americaninequality.substack.com
The 2nd largest police force in the country has added a controversial new tool to its arsenal. The Chicago Police Force announced on September 2, 2020 that it was partnering with Amazon Ring for a new neighborhood surveillance program. The CPD joins 1,694 police departments across the country that have created similar partnerships, hoping that Amazon can help these communities reduce crime and increase security.
However, state and local governments have had to navigate the uncharted waters of tech governance in policing, particularly around facial recognition. Although more than half of all Americans are in a facial recognition database, this space is currently not well regulated. Many fear that profiles will be created on them without their consent. This is particularly dangerous for black communities that already endure increased threats from police presence. All eyes turn to Chicago’s Chief Information Security Office to create a path forward that upholds privacy, transparency, safety, and equitable delivery of public services.
Context and Analysis
Ring smart doorbells are equipped with video camera surveillance. When Ring cameras detect motion, they start recording whatever is in view whether it’s a burglar, passerby, or Amazon delivery person. This video is sent to an app on the homeowner’s phone, which can then be shared publicly or privately with direct messages. Amazon has kept Ring as a separate entity since its acquisition of the company in 2018 for $1.2 billion. This is important, since Amazon’s Rekognition business line has facial recognition technology that could be coupled with Ring surveillance.
Amazon Ring-Police partnerships are on the rise. The first Ring-Police partnership was announced on March 22, 2018 with the Greenfield, Wisconsin Police Department. Since then, Amazon has received 5,915 requests for video surveillance footage from the 1,694 police departments. The Chicago Police department announced a partnership with Ring on September 2, 2020 and 5 police departments in neighboring regions also have partnerships. There are 180 Ring-Police partnerships across Illinois.
Racial tensions are also on the rise, especially in Chicago. As crime and homicides have increased 50% since January, up from 392 to 588 this year, the city’s police presence has intensified. Since the 2014 killing of Laquan McDonald in Chicago, race-focused protests have dominated the city almost continuously, with two incidents in Pilsen and Little Village in the last 6 months alone. Chicago’s CIO must keep these racial tensions in the foreground in order to equitably and appropriately navigate new police tools.
Several cities across America have banned facial recognition due to its untested governance and controls. San Francisco, Boston, and Portland are among dozens of cities that have banned this technology, citing risks of “biases against Black people, women, and older people.” Research from Joy Buolamwini at the MIT Media Lab has shown that facial recognition softwares have error rates of 34% in darker skinned women, often because the underlying databases that train the algorithms are 77% male and 83% white.
The City of Chicago must consider 5 primary stakeholders and 4 secondary stakeholders as it considers implications for facial recognition software. First, governments must uphold their contracts with citizens to provide safety and public services. Second, police are the enforcement arm of the city and the government must ensure that police are properly equipped to defend and protect. Third, politicians want to champion wins and deliver on campaign promises to improve their cities. Fourth, criminals and burglars are stakeholders since they may have either been driven to a life-of-crime (often as a last resort) or would benefit from seeing crime rise so that they could hide amongst the masses. Fifth, Amazon is a stakeholder since it wants to work more closely with local governments and receive funding for its services.
The secondary stakeholders must also be considered in making any economic decision. The Mayor’s office is just one part of larger complex of government services ranging from courts, to rehabilitation clinics, to prisons, all of which depend on proper social safeguards. Competitors are interested in seeing Amazon fail and taking over their position to win lucrative government contracts. Tax collectors and government budgeters will want to ensure that cities can pay for new policies. Lastly, neighborhoods value preserving their identity, their togetherness, and their security in ways that can often be quite different from the values of citizens.
The Chicago CIO has a duty “to successfully transform digital services and build trust between residents and government.” The city has a long history of performing digital services. Danielle DeMerer, the city’s former CIO, said that “In 1999, Chicago was one of the first cities in the US to implement a 3–1–1 non-emergency management system to make it easier for residents to access city services and offload calls from the city’s 9–1–1 emergency management system.” The city’s emergency management system has once again become overloaded and new digital services are needed.
The CIO’s first priority is thus to improve government services through digital means. This includes reducing costs, increasing efficiency, improving access, and limiting waste through digital tools. In this case, the CIO must evaluate how it can improve emergency services by digitizing parts of the criminal justice process.
The CIO’s second priority is to uphold citizen privacy in any new initiative. Privacy at the core of user trust. Without privacy, citizens will lose trust in their governments and use of public services will break down. In this case, the CIO must ensure that citizens know how their data is being captured, used, and transferred and who has access to that information. Accidental release of private Ring footage would be horrific for the CIO’s office.
The CIO’s third priority is to uphold transparency in any new initiative. This means that the public must understand why a program is occurring, how it is being executed, and how it will impact people’s lives. Project transparency will often fall across multiple domains including legal, budgetary, and regulatory affairs. In this case, the CIO must show how the Chicago police department is encrypting, securing, and deleting Ring video footage against breaches as well as deploying facial recognition in videos.
The CIO’s final priority is to ensure equitable delivery of digital public services. One community, neighborhood, race, or socioeconomic cluster of people cannot receive unequal treatment under the law and digital services are no exception. As discussed in greater detail below, facial recognition has possibly much more pernicious effects for black citizens. In this case, the CIO must ensure that black Chicagoans or lower-income Chicagoans do not see negative externalities as a result of this digital policy.
While the CIO should focus on data security, the CIO does not have to focus as deeply on community security. This is the priority of the Mayor, the Chief of Police, and the Chicago Attorney General. The CIO’s responsibility is to improve government digital services.
The Path Forward
This paper does not recommend that the Chicago CIO pursue a facial recognition partnership with Amazon-Ring. The risks of increasing racial bias, of creating irreversible cybersecurity threats, of creating a state of constant fear and surveillance is too great a burden for Chicago to bear. The Ring-CPD partnership is already in place, the city does not need to go further in building out facial recognition capabilities. The rationale below explains how this conclusion was reached.
Law enforcement officials argue that there are two pros to using facial recognition technology — improved efficiency and improved optionality.
- Improved efficiency — The Department of Homeland Security has argued that facial recognition technology can help the government more quickly screen passengers at airports and process people at immigration checkpoints. This process is currently very manual, and often requires a federal agent holding up a picture of your face (either a passport or driver’s license) next to your face and looking back-and-forth to see if they are similar. In other cases, humans search through massive databases of profiles to find a match, making sure that an individual is not on a banned list. This tedious human process could be sped up and improved with machine capabilities, saving time and money.
- Improved optionality — In 2019, New York’s police commissioner James O’Neill wrote an Op-Ed in the New York Times arguing that “facial recognition makes you safer.” After explaining that he used to use ink-finger print cards and Polaroid cameras for mugshots, he goes on to say that facial recognition is just the next step in a long-line of evolution of police tools. “Technology has improved the profession beyond what the most imaginative officer could have conceived in those days. These innovations include facial recognition software, which has proved its worth as a crime-fighting resource since we adopted it in 2011… We use these methods solely to fill in missing or distorted data.”
Policymakers argue that there are two cons to using facial recognition technology — increased racial bias and increased cybersecurity and governance risk. A greater discussion of the cons of costs are outlined below.
- Increased racial bias — Facial recognition technology has a racism problem. The National Institute of Standards Technology (NIST) analyzed 189 facial recognition algorithm from 99 developers using 18.3M images of 8.5M people were 10x to 100x more likely to misidentify African Americans, Native Americans, and Asian Americans compared to white faces. Deploying facial recognition technology with the Amazon-Ring cameras may exacerbated racial tensions that are already at a boiling point in Chicago. Police may wrongfully arrest more black Chicagoans and civilians may experience a greater culture of fear and suspicion, that often seems into xenophobia and racism, according to the ACLU.
- Increased cybersecurity and governance risk — The threat assessment of biometric databases is incredibly high. While you can always change your password or pin if you are hacked, you can’t change your face, your fingerprint, or your iris. Facial recognition databases have been hacked in the past. In 2019, hackers stole photos of 100,000 travelers’ faces and license plates from a 3rd party surveillance company, Perceptics, that was working with the Customers and Border Protection agency. Even beyond the breach of privacy, this poses additional long-term risks since hackers can use these images to create fake identities of people online, spreading even more fear online.
Costs from implementing (or not implementing) facial recognition software in policing
The costs associated with not pursuing a facial recognition strategy emerge from 2 areas — the cost of lost efficiency and the cost of unsolved crimes.
- In a small footnote in its November 2020 report, the Customs and Border Patrol has said that for Global Entry terminals, “Facial recognition transactions have reduced kiosk processing time, already short, by an astonishing 90 percent.” This rate of efficiency could save millions for the CPD.
- The costs of unsolved crimes remains high, which could decrease with facial recognition. According to researchers at Iowa State, every murder costs society $17.3M, every rape costs society $448K, every armed robbery costs $335K, and every aggravated assault costs $145K. While these numbers are controversial, they may shed light on possible price savings for crime reduction. However, solving crimes also has a racial bias problem, which could be exacerbated by facial recognition. In evaluating 849 murders between January 2018 and July 2020, in Illinois, 47% of the cases were solved during those same 19 months when victims were white, 33% when the victim was Hispanic, but only 22% when the victim was black. Facial recognition may be good for solving white crimes, but not others.
The costs associated with pursuing a facial recognition strategy emerge from 3 areas — the cost of paying for the service, the cost of improving IT resiliency, and the increased costs of litigation.
- The costs associated with paying for Amazon’s facial recognition software are remarkably low — the Washington County Sheriff’s Office in Hillsboro Oregon pays just $7 per month for the service. Nevertheless, the 2020 police budget included $600,000 for “smart policing technologies”.
- The costs associated with improving IT resiliency may be very high. In 2012, CPD requested $2 million to support real-time video-based face recognition. They also bought an exceptionally large amount of network hardware and spent $450,000 on enterprise-class database and computing infrastructure. The Chicago police department’s budget was $1.65B in 2019 and grew to $1.78B in 2020, with technology accounting for 21% of costs or $378M. Part of this budget increase went to encrypting new police radios, which was already the second time the police had to do this. More sophisticated technologies may prove to be even more expensive to encrypt and maintain sufficient IT protocols.
- The costs of litigation are expected to be high, though many government lawsuits are still pending. In 2018, Chicago spent $113M on police misconduct lawsuits while Facebook had to pay $650M for its improper use of facial recognition software, which would account for 1/3 of the total CPD budget. The Illinois Biometric Information Privacy Act (740 ILCS 14/1) passed in 2008 may cause significant legal challenges. In 2019, the Illinois Supreme Court ruled that individuals do not need to show they suffered harm other than a violation of the law to bring a suit — “As a result, entities are at greater risk of liability for failure to follow legally required procedures for handling biometric information collected or stored in Illinois.”
Hypotheses to Test
While the recommendation for the CIO is not to implement facial recognition software, the CPD may want to test whether crime is reduced in similar cities across America that are using facial recognition. The map below highlights the regions where police departments are using facial recognition software.
This paper recommends evaluating the success of the Los Angeles police in using facial recognition software. LA is the next largest police department in America after Chicago and uses facial recognition software for crime identification. CPD can evaluate LA as a test-case to determine how effective facial recognition is in reducing crime. Amazon-Ring alone does not
The CIO may also want to conduct a survey with Chicagoans to evaluate their comfort with facial recognition. Surveys currently show that only 16% of Americans are comfortable with facial recognition software, but with crime on the rise in Chicago, this stat may jump higher.
The CIO has few implementations steps because the recommendation is not to pursue facial recognition, however, the CIO will need to implement a communication plan for this decision. The CIO should work with the Mayor’s communication department and the CPD communication department to publicly announce that Chicago is not pursuing facial recognition. This is critical for creating trust with citizens, particularly as several cities and municipalities have clandestinely rolled out programs without full citizen awareness. The city should stress that it found that facial recognition can be “psychologically unhealthy”, has shown to promote racial bias, and requires new IT safeguards and governance before it can be safely implemented.