By: Ángel Díaz
In every age, police forces gain access to new tools and technologies that may advance their mission to prevent and combat crime. The deployment of new technologies requires an understanding of their impacts on the fundamental rights of the communities that police serve and the development of safeguards to prevent abuse. The New York Police Department (NYPD), however, has purchased and used new surveillance technologies while attempting to keep the public and the City Council in the dark. This chart provides an overview of the NYPD’s surveillance technology, based on publicly available information, as well as the potential impact of the use of these tools.
Because the police insist on complete secrecy, however, the picture is far from complete. The NYPD should not be allowed to prevent the public and its elected representatives from learning basic information necessary on these technologies, which is critical to effective oversight and the establishment of safeguards to protect the privacy and civil liberties of New Yorkers. The POST Act, introduced by Council Member Vanessa Gibson and currently supported by 28 co-sponsors, would require NYPD to take these steps.
Facial Recognition
How It Works:
Facial recognition systems attempt to identify or verify the identity of individuals based on their face. Different systems analyze face characteristics in photos or video feeds, or through real-time surveillance.
Impact:
Facial recognition raises the following concerns: Race, Gender, and Age Bias. Numerous studies have found that facial recognition performs poorly when analyzing the faces of women, children, and people with darker skin tones. 1 This places communities already subject to over policing at greater risk of misidentification. Privacy. Facial recognition is recognized as extraordinarily intrusive, challenging reasonable expectations of privacy and lacking necessary oversight. This is why a number of groups have called for a moratorium on facial recognition. Free Speech. Law enforcement use of facial recognition can chill the exercise of First Amendment rights by exposing protesters to persistent surveillance and identification. Regulation. There have been widespread calls for its regulation2, and some cities — such as San Francisco3; Oakland4, CA; and Somerville, MA5 — have even banned its use.
NYPD Policy & Scope of Use:
NYPD’s Facial Identification Section (FIS) runs static photos obtained from various sources, including databases of arrest photos, juvenile arrest photos of children as young as 11, and photos connected to pistol permits, among others.6 The system analyzes a photo against those databases and generates potential matches.7 The system will return a list of 200+ potential matches from which an FIS investigator selects one.8 Where the footage is blurry or otherwise unusable, the NYPD can use photo editing tools to replace facial features in a reference photo so it more closely resembles those in mugshots.9 The NYPD has also run photos of celebrities through its facial recognition system to try to identify suspects that resemble the celebrity where the original photo returned no matches.10 The effectiveness of these techniques is doubtful.
Video Analytics
How It Works:
These systems analyze surveillance camera footage and attempt to isolate people and objects within the video feed. Video analytics use algorithms to spot particular articles of clothing and luggage. Certain versions claim they can find people in surveillance footage that match a particular hair color, facial hair, and even skin tone.
Impact:
Video analytics raise the following concerns: False Positives. Information from video analytics can be incorrect and lead to unnecessary and potentially dangerous police encounters. Free Speech. Video analytics, like facial recognition, can chill First Amendment activity by exposing individuals to persistent surveillance as they move about the city. Racial Bias. Without adequate controls, targeting individuals based on their perceived ethnicity has the ability to exasperbate racial disparities in policing. Privacy. Video analytics allow for persistent surveillance as individuals move throughout the city, challenging traditional expectations of privacy
NYPD Policy & Scope of Use:
No standalone NYPD policy is available, though video analytics may fall under the Public Security Privacy Guidelines that govern the NYPD’s Domain Awareness System. These guidelines make no mention of video analytics, however, and they do not include standards governing the use or storage of analytics information. IBM developed object identification technology through a partnership with the police that gave the company access to the department’s camera footage.11 The NYPD then acquired IBM’s object identification system to incorporate it into the NYPD’s Domain Awareness System. 12 As of April 23, 2019, IBM stopped marketing certain versions of its Video Analytics program to additional cities.13 It is not clear what this means for IBM’s existing customers. According to the NYPD, the analytics system is intended to automatically alert NYPD officials to activities, such as “suspicious package was left” or “loitering.”14 A version of IBM’s Intelligent Video Analytics 2.0, which allows users to search based on ethnicity tags, was allegedly tested but never incorporated into the NYPD’s broader surveillance infrastructure.15
Automated License Plate Readers
How It Works:
Automated license plate readers (ALPRs) are devices that are attached to police cars or fixed on poles to capture the license plates of all cars passing by. License plate reads are also frequently run against a “hot list” of, for instance, stolen cars or AMBER Alerts. In addition to license plates, ALPRs can capture photographs of cars, along with photos of the driver and passengers. This information is uploaded to a database where it can be analyzed to study movements, associations, and relationships to crimes.
Impact:
ALPRs raise the following concerns: False Positives. Information from ALPRs can be incorrect and lead to unnecessary and potentially dangerous police encounters. Privacy. ALPR data can provide a detailed account of an individual’s movements. It can be used to target people who visit sensitive places, such as immigration clinics, protests, or houses of worship. Impact on Immigration Status. Police agencies can choose to share their ALPR information with federal immigration authorities. According to a public records request, ICE has received ALPR data from 80 different police departments, including Fairfield, CT; San Diego, CA; Orange County, Texas; and Athens-Clarke County, GA; among others.34 It is not known whether the NYPD shares ALPR data with ICE, but the Public Security Privacy Guidelines permit the sharing of ALPR information with government entities.
NYPD Policy & Scope of Use:
The NYPD operates nearly 500 license plate readers as part of its Domain Awareness System,35 and as of 2013, the department had a database of 16 million license plate reads.36 The NYPD has used license plate readers to collect information about the cars parked in mosque parking lots.37 Through its contract with the vendor Vigilant Solutions, the NYPD now has access to a database that contains over 2.2 billion license plate reads.38 Vigilant Solutions has a national database of license plates, a national network of private ALPRs, and analytical tools that allow police to “stake out” areas, predict where certain individuals may be, and track individuals outside of New York City.39 We do not currently know if NYPD shares the data it gets from its own ALPRs with other clients of Vigilant Solutions as well as other law enforcement or federal immigration agencies, as some cities do.
Drones
How It Works:
Drones are remotely operated aircraft — ranging in size — that can be equipped with various cameras, sensors, and other devices. For example, they can deploy cameras capable of facial recognition, and can also contain GPS trackers and Stingray devices.
Impact:
Drones raise the following concerns: Privacy. Without proper oversight, drones can engage in forms of surveillance that can redefine reasonable expectations of privacy. Drones can also be used to collect information about bystanders who are not connected to a law enforcement investigation. These risks are largely invisible, as drones can be difficult for ordinary persons to detect or protect against depending on their size or altitude. Free Speech. Without proper oversight, drones can be deployed to surveill individuals in ways that chill free expression
NYPD Policy & Scope of Use:
The NYPD’s policy specifies that it will not equip drones with facial recognition, but it contains a large carve-out for situations where there is a “public safety concern.”49 It is unclear if there are any restrictions on running historical drone footage through a separate facial recognition system. The policy also specifies that drone footage will only be retained for 30 days, but it contains a carve-out that allows this period to be extended for various types of legal investigations.50 According to the NYPD, the department deploys drones for uses such as crowd control, hostage situations, and reaching remote areas. The NYPD says drones will not be used for routine police patrols, to enforce traffic laws, or for “unlawful surveillance,51 but the NYPD has deployed drones to monitor protesters at least once during the 2019 NYC Pride March.52
Body Cameras
How It Works:
Body cameras are used to record an officer’s interactions with the public and store the video for future review or use in criminal or civil proceedings. While body cameras have been promoted as a tool for police accountability, they have largely functioned as evidence-gathering devices.
Impact:
Body cameras raise the following concerns: Effectiveness. As part of the settlement related to the NYPD’s unconstitutional stop-and-frisk program, a federal judge ordered the NYPD to develop a mechanism for officers to electronically record certain police encounters.61 However, the cameras remain under the control of police, who can decide when to activate them. Even when the cameras are rolling, police officers can add audio commentary that skews public perception of an incident (e.g. yelling “stop resisting” to a cooperating person). Privacy. Absent safeguards, body cameras can function as mobile surveillance devices, recording information about people and places that officers encounter while on patrol, regardless of their relationship to a suspected crime. Future iterations of body cameras may be equipped with facial recognition technology,62 raising additional concerns about privacy, effectiveness, and racial bias.
NYPD Policy & Scope of Use:
In New York City, members of the public can request video under the Freedom of Information Act, but when it relates to evidence in a criminal case the video is turned over to the prosecutor’s office. If a camera records an officer-involved shooting or other high-profile incident, NYPD works with “relevant authorities” to determine if video can be made public.64
SkyWatch & TerraHawk Surveillance Towers
How It Works:
Surveillance towers allow officers to monitor areas from several stories above street level as well as record movements within a targeted area. Each SkyWatch tower contains flood lights, a command desk, devices to detect vehicle speeds, tinted windows, digital video recorders, and customized surveillance cameras.65 The standard equipment placed on TerraHawk towers is unknown, but their patented technology contemplates the use of surveillance cameras along with infrared detectors, motion detectors, and a thermal imaging device.66
Impact:
Surveillance towers raise the following concerns: Privacy. Surveillance towers impose a feeling of persistent monitoring, challenging reasonable expectations of privacy. Surveillance towers can also be used to collect information about bystanders who are not connected to a law enforcement investigation. Free Speech. Persistent monitoring from surveillance towers can chill associations among individuals.
NYPD Policy & Scope of Use:
In New York City, members of the public can request video under the Freedom of Information Act, but when it relates NYPD may deploy surveillance towers in response to a rise in crime within a particular area,67 but they have also been used to monitor protests, such as Occupy Wall Street.68 The current number of towers deployed by NYPD is unknown. Surveillance towers are also used to collect “probative” and “potentially probative” images, according to patrol guides, but the meaning of these terms is unclear. According to media reports, TerraHawk Towers have been deployed in Staten Island, Far Rockaway, Coney Island, and Howard Beach. 69 SkyWatch have also been deployed in Harlem70, Crown Heights71, downtown Manhattan (Zuccotti Park)72, Bedford-Stuyvesant Brooklyn73, and the Lower East Side of Manhattan (Tompkins Square Park)74.
Source: https://www.brennancenter.org/sites/default/files/2019-10/2019_10_LNS_%28NYPD%29Surveillance_Final.pdf
