If Not Reviewed Properly, Bodycam Footage Is an Expensive Wasted Opportunity

For more than a decade now, many police departments in the United Kingdom and the United States have been using body cams as way to improving trust with the public police by providing transparency and accountability.

June 21, 2024 by Enrique Dans Leave a Comment

For more than a decade now, many police departments in the United Kingdom and the United States have been using body cams as way to improving trust with the public police by providing transparency and accountability.

The devices used for this have evolved over time and have become a hub for permanent two-way communication, capturing high-definition video through one or two cameras (often one is worn on the chest, with a second point of view that is fixed to the head), as well as audio, precise location data, etc. This generates, when multiplied by the number of active police officers, huge amounts of video that must be stored and processed. In many cases, cameras and their cloud storage have become the most significant cost for police departments, with annual bills exceeding several million dollars.

However, a paradox arises: when a single police department accumulates several million hours of high-definition film, how can that footage be processed to find some kind of useful information? The answer seems clear: only interactions that really become a problem are reviewed, while the vast number of them are simply stored without anyone ever being looked at. In many cases, the families of victims of police violence have to fight in court to access body cam recordings of the police officers who were present, and they have no guarantee that these recordings will be adequately preserved.

To avoid this and to get a better return on the investment, some police departments are doing something that seems, in principle, paradoxical: increasing the bill they pay in technology by including AI packages able to review huge amounts of material to monitor anything from the use of force, to harassment, inappropriate vocabulary or interactions with the camera itself to turn it off or interrupt the sound recording temporarily.

One company, Truleo has created an algorithmic tool that can review footage and turn it into actionable and useful information, able to identify possible problematic behavior so it can be identified and associated with specific police officers. The cost of the tool, around $50,000 a year, is an increase from a budget considered by many to be very high, but it could be the way to get the most from body cams. This is a perfect example of how, on many occasions, an initial investment in technology requires further spending to solve problems that were not initially detected, and how these additional investments, if not made, reduce the value of the original outlay.

In this case, we are talking about a fundamental requirement: the police have, in a democracy, a monopoly on the use of violence, but it is essential that it be exercised with proportionality and responsibility, without excesses or abuses of authority. Cameras are a fundamental element of control, but the huge amount of material they generate and the need to process it properly can soon become an impediment.

However, police departments have in their hands not only those hours and hours of video, but also the ability to label problematic behaviors and use them for algorithm training, in the same way that YouTube, Instagram or TikTok do. Another of the many cases we are seeing where algorithms can prove their value.

(En español, aquí)

—

This post was previously published on MEDIUM.COM.

***

—

Photo credit: iStock.com