Privacy Concerns to Address if Government Expands Use of Body Cameras
Written by Jake Laperruque
This week, the Center for Democracy & Technology submitted a set of recommendations to the President’s Task Force on 21st Century Policing outlining key protections for the use of body cameras. While certainly not a cure-all for police misconduct issues, we believe body cameras have strong potential to aid civilians and improve community-police relations; pilot programs have reduced citizen complaints by as much as 88 percent and decreased police use of force by 60 percent.
However, these encouraging statistics and urgent calls for reform threaten to cause a hurried rollout of body cameras without proper consideration of privacy protections. Pilot programs are rapidly being enacted through the country, including in New York, Chicago, Washington, DC, and Los Angeles. The President recently proposed providing funding for 50,000 new body cameras, and Congressional legislation has been introduced making federal funding for law enforcement contingent upon body camera use.
Body cameras offer great potential benefits, but also represent a powerful new technology that could be co-opted as tool for mass surveillance.
Body cameras offer great potential benefits, but also represent a powerful new technology that could be co-opted as tool for mass surveillance. Their use should be paired with firm guidelines to protect privacy, and ensure that body cameras are a check on government power, rather than a means of improperly augmenting it.
We support the following guidelines for body cameras, and believe federal funding for camera use should be tied to both best practices and robust involvement from local communities controlling how camera technology is deployed and used.
1) Requirements and Limits for Recording:
The most fundamental policy question is when cameras should be on. This policy should strive to make sure any potential interaction that could result in misconduct or a complaint is recorded, but also account for privacy interests of both officers and civilians, and questions of practicality. Research demonstrates that providing greater discretion results in a substantial decrease in recordings. Therefore, a model general recording requirement should limit discretion, and require cameras generally be turned on whenever officers are interacting with the public.
However, a broad general recording policy should be paired with strong exceptions to compensate for privacy needs. An effective means of addressing this may be to require that cameras be turned off in locations where civilians have a reasonable expectation of privacy, such as bathrooms, dressing rooms, and locker rooms. Policies should also account for issues of practicality, using rules such as requirements that cameras are equipped with a “pre-event video buffer” and standardized rules for switching batteries and maintenance checks of cameras.
2) Rules for Notification and Opting-Out:
Individuals should be given as much control as possible in regulating when recording is directed at them, and the ability to “opt out” as desired. This control will promote individual privacy from unwanted government surveillance, but also facilitate police interviews with individuals – notably crime victims and confidential informants – that may be reluctant to discuss an investigation on video.
In order to achieve this goal, when effective notice is feasible, some form of notification should be required that cameras are recording. Quite simply, individuals cannot exercise a right to opt out of being recorded if they do not know a video feed is on. Notification could occur via an explicit statement, or through indirect notification such as a “recording” light accompanying a camera.
3) Requirements and Limits on Retention:
Retention limits are a crucial backend method of ensuring that body cameras serve as an oversight tool, and prevent creeping use of cameras for dragnet surveillance. However, these limits must account for factors such as the length of time that may elapse before a civilian files a complaint, evidentiary use, and examination of evidence by criminal defendants. As an overall baseline, departments should be required to retain all video for the length of time civilians may file complaints.
Evidentiary video – video consisting of “an incident or encounter that could prove useful for investigative purposes, such as a crime, an arrest or citation, a search, a use of force incident, or a confrontational encounter with a member of the public” – has added value for civilian complaints, criminal investigations, and criminal defendants. However, the range of criminal offenses should be considered in establishing retention periods and balancing the seriousness of a crime against privacy concerns. State evidentiary rules could serve as an effective foundation for reaching this balance, where the retention period for evidence about minor crimes is limited, but evidence about more serious crimes – such as homicides – can be much longer, and sometimes are indefinite. At a minimum, evidentiary video should be retained until the conclusion of a trial or the close of the investigation to permit review by potential criminal defendants.
But evidentiary video is only a tiny percentage of the video footage that will be collected by cameras. It is crucial that most footage be deleted as soon as the time limit for complaints has passed – reducing the risk that footage will be deployed to monitor innocent individuals not involved in any wrongdoing.
While video is being retained, access should be limited. Video should be generally inaccessible, unless needed as evidence or for internal investigations. Strict prohibitions should exist against any editing of video. To prevent retroactive development of justifications for police actions such as arrests and police stops, officers should not have access to their own feed prior to filing of reports. High standards for data security should exist, regardless of whether videos are stored directly or with a third party.
4) Limits and Protections Regarding Dissemination:
While the added police accountability that body cameras promise can only be achieved if a video feed can be released to affected civilians and other parties devoted to police oversight, dissemination rules must account for the privacy interests of those being recorded. While civilians alleging police misconduct will have the most direct interest in obtaining video from body cameras, other parties – such as civil rights groups, government transparency groups, and media – will also have a legitimate interest in obtaining video feed for legitimate public interest goals.
However, if video feed is to be made generally available, precautions should be put in place to protect the privacy of individuals recorded, especially given potential sensitivity of police interactions. Body camera feeds should be redacted in some circumstances to block 1) any personally identifiable information and 2) video whose disclosure constitutes a clearly unwarranted invasion of personal privacy. While redaction efforts will require time and resources, tools such as face blurring technology can make the process significantly easier, and are already being employed by some departments.
5) Limits on Use of Facial Recognition:
Use of facial recognition in combination with body cameras represents a significant risk to privacy, and should be barred or sharply limited. Development of face prints from body camera feeds represents a significant threat to privacy and activities protected by the First Amendment. Face prints could be cataloged from officers’ recordings of religious ceremonies, political rallies, or public protests.
Allowing officers to run existing face prints against video that is being recorded from body cameras creates risks to privacy as well. This practice could be used to locate and monitor the activities of individuals not suspected of wrongdoing, especially in cities with large police forces. Courts and state legislatures are increasingly recognizing the privacy value of location information – including in public – and establishing warrant-for-location requirements. Use of facial recognition in combination with body cameras should not serve as a loophole for these protections.