Many law enforcement agencies have procured or are considering the acquisition of body-worn video cameras designed to document interactions with the public. Cities and counties throughout the country have decided to invest in these devices and the associated storage required primarily based on the premise that these cameras will document any evidence of the use of excessive force by law enforcement officers. The assumptions made about the use of the BWCs is that their presence will deter the use of excessive force but in the event that deterrence fails there will be evidence of misbehavior that can be used to achieve the ends of justice.
Police officers and agencies have been open and in many cases encouraging about the decisions of their communities to acquire these capabilities. The thinking seems to have been that the police believe the very act of recording people in their interaction with police will influence better behavior and less unwarranted complaints about police behavior.
The combination of pressure from community groups along with at least acquiescence from police has resulted in the adoption of these technologies in well over half of the police departments in the U.S. The trend continues as more and more cities and counties decide to adopt body-worn cameras, even though the research conducted to date gives rise to questions about the extent to which these devices achieve the desired objectives. In a review of the research conducted regarding the effects of the adoption of body-worn cameras in policing, Cynthia Lum and her colleagues discovered that the impact is not what advocates expected from these devices. [1] The evidence of reducing the police use of force is mixed in the 70 studies reviewed, and the effects on the public are also not as clear as proponents might have desired.
But there is another aspect of the impact of these devices that has been often overlooked in the analysis leading to the decision to deploy these devices. When policies about when to make recordings are written to ensure that officers turn on and keep on the recording, a substantial amount of video recording is generated. Estimates are that 4 hours of video recordings are generated in a single shift. The amount of video generated is overwhelming. What many jurisdictions fail to take into account is that some percentage of this video material becomes evidence when it is associated with an arrest and subsequent trial. And when this mountain of evidence is forwarded to the prosecutor along with the case records, the prosecutor is obligated under federal and state laws to review the evidence to provide any exculpatory information to the defense attorney and also to redact any personal privacy information that is not relevant so as to protect the privacy and civil liberties of people who may have inadvertently be captured on the video.
The cost impact of the obligation to review and redact the video from body-worn cameras creates a very unexpected weight of evidence that is formidable. There is a general finding that for every 50-100 cameras deployed, the jurisdiction doing so will need to hire one additional attorney. In Virginia, the commonwealth has justified paying for one additional attorney for every 75 cameras deployed in local jurisdictions.
Most jurisdictions have embarked on the deployment of body-worn cameras without taking into consideration the costs of dealing with the evidence generated by these devices, and the obligations of the criminal justice system to properly deal with the evidence that is generated. One of the challenges that must be faced is that in many jurisdictions the criminal justice agencies are quite separate from the law enforcement agencies in terms of budget coherence and organizational decision-making. When a city within a county makes the decision to proceed, in many cases the obligation and the cost impact falls on the county where the prosecutor agency exists.
A more unified consideration of the costs can be made in consolidated jurisdictions. In the metropolitan area of Nashville and Davidson County, the mayor and council are faced with both the police costs and the criminal justice system implications, and the District Attorney General there was sufficiently concerned about the potential implications of body-worn cameras on the cost of the required evidentiary response that he commissioned a study to develop a life-cycle cost model that could be used as a framework for making policy decisions on how to handle the potentially overwhelming amount of evidence that would be generated from the deployment of over 3,000 cameras. The study and the cost model were assigned to Wormeli Consulting and the report can be found below.
The outcome of this study was the decision to implement body-worn cameras in Nashville incrementally, so as to enable the collection of the data from early deployments that would drive a more informed set of policies on how to handle this evidence and still meet the obligations that the criminal justice system has to protect privacy and ensure the fair administration of justice.
Disruptive and innovative new technologies often come with unforeseen consequences that revise the economic as well as the operational assumptions on which their deployment was based. It is always helpful to think through the operational concept in detail so as to understand the implications of a paradigm change that such new technology brings.
[1] Lum, C, Stoltz, M, Koper, CS, Scherer, JA. Research on body‐worn cameras: What we know, what we need to know. Criminology & Public Policy. 2019; 18: 93– 118. https://doi.org/10.1111/1745-9133.12412
Some very good points. The other positive outcome is deterance of both the officer and the subject to react in a violent manner, the ability to use the video for training purposes and assisting is teaching de-escalation approach’s.
The challenge is having the Sgt. review the volume of video produced in a shift. In Newark the officers generate 4-6 hours of video consistent with the Consent Decree Guidelines. The major incidents like Use of Force, arrest, search etc. are flagged for review. The other 80% are not. Within the 80% there may be officer safety, un intended bias, poor community relations that need to be identified.
We are working with NJIT, NPD, IBM and Rutgers to develop a filter that uses AI to idea these type of behaviors. It is based on the machine learning from Watson. Let me know if you want to discuss.
Tom O’Reilly