Calls for police officers to wear body cameras at all times have skyrocketed after Michael Brown was shot and killed by police officer Darren Wilson under contested circumstances.
Given that encounters with the police can sometimes be lethal affairs--particularly for blacks and Latinos--body cameras seem like a no-brainer solution to remove uncertainty from those incidents. Yesterday we looked at some of the potential privacy concerns around police body cameras. Today we’re going to take you through the actual hardware and software of these wearable gadgets.
Two of the largest companies selling these cameras in the U.S. are Vievu and Taser International--the latter is the maker of the eponymous Taser stun gun. Vievu president Steve Lovell, a former marine and police officer, says that their product can “take a jury into the crime scene, when otherwise they would only read a police report.”
Rick Smith, the founder and CEO of Taser, says that his company's “100-year vision is to make obsolete the idea of shooting someone with a bullet” and that Taser got into making cameras to provide transparency into violent confrontations “so that everyone can know what happened.”
A widely cited study of body cameras used by police in Rialto, California shows a drop in both complaints against officers and police use of force. While that's a net good in and of itself, cities around the country are now also seeing this as a way to save money in the long term. Both camera manufacturers and civil liberties group advocating for body camera adoption believe that the cost savings from investigating fewer complaints of police misconduct will make the cameras pay for themselves in short order.
As a result, both Vievu and Taser have seen interest in their products shoot up rapidly. Police chiefs have been keeping their phones ringing off the hook these past couple weeks.
Taser's AXON body camera and Vievu's LE3 camera systems are fairly similar, with some notable differences. Taser also offers a model called AXON Flex, which can be mounted on an officer's sunglasses, collar, or hat. Vievu's cameras are manufactured overseas, while Taser has parts sourced globally but assembles its devices in the U.S.
As police generate recordings throughout the day, videos are stored locally on the devices until the officers return to headquarters and download the files into their department's database. At the end of a shift with a Vievu LE3 camera, an officer removes her camera and connects it to a computer with a USB cable. Using Vievu's accompanying proprietary software, the officer then transfers the videos into their department's system and write up notes on events from the day.
With the AXON models, the process is a little more seamless. Using Taser's accompanying smartphone app, officers can tag and categorize their incidents back in the squad car immediately after they happen. When her shift ends, an officer just drops the camera into a cradle that acts as charger and automatic sync system.
Both camera systems are integrated with software services for storing and managing the video. Vievu's offering is called Veripatrol, while Taser's is Evidence.com. Both provide fine-grained permissions so that certain files can be restricted and any time a user accesses a file is tracked. Both encrypt the videos in such a way that even Vievu and Taser cannot access them, unless invited to do so by a department. Most importantly, both provide simple, scalable solutions that take the burden off of managing big data for police departments.
“Your average police department in the United States is a city agency with less than 50 people,” says Smith. “Most police departments do not have a full-time information security officer.”
Evidence.com is offered on a software as a service model, where departments pay a monthly subscription fee to use the cloud storage service, a la Dropbox. Veripatrol on the other hand is bundled for free with the Vievu cameras and police departments can run the software locally on servers in their own department. If a department prefers cloud functionality, Vievu offers Veripatrol as a managed service hosted on Amazon Web Services. In that case the department would only pay for its actual hosting costs on Amazon's cloud.
As Evidence.com is also hosted on Amazon Web Services, both Smith and Lovell are quick to point out that the CIA recently announced a deal for Amazon to manage data for intelligence community.
“When we started in this space five years ago and talked about putting law enforcement data in the cloud, there was a very skeptical response,” says Smith. Lovell says that Vievu encountered a similar hesitance at first, saying that “law enforcement wanted the evidence in their basement.”
“But we've seen that absolutely turn on its head with the CIA choosing Amazon to build them a data center,” says Smith. “It was a seismic shift. When police ask about security and we point out the CIA's data center, it's basically the end of the conversation.”
In either instance, one potential concern is vendor lock-in. Smith himself even makes the comparison to the Apple ecosystem of iPods and iTunes, which, while convenient, has historically been criticized for using non-standard components and being hard to break free of once bought into. The ease of changing camera providers and software systems is something cities need to be wary of when contracting with a camera vendor.
But as ACLU senior policy analyst Jay Stanley points out, “the risk of vendor lock-in depends on how much need a department has for past video.” And Stanley's position is that video should not be kept any longer than it absolutely needs to.
“We've called for recordings to be disposed of in a reasonable period of time unless for some reason they are flagged,” he says. “They could be flagged either because there's been a formal or informal complaint of abuse against an officer or because police believe it to contain evidence of a crime. In the absence of one of those flags, video should never see the light of day. We consider a reasonable period of time to be months or weeks, not years.”
One key difference between the Taser and Vievu models is in how much and how well they film. The AXON body camera boasts a wide-angle 130 degree field of view compared to the LE3’s 68 degrees. However, Vievu's cameras record in 1280x720 HD while Taser's captures in 640x480.
“We have worked with video forensic experts to try to create a very flat image so if forensic enhancement is required we don't want any distortion from the camera's video segment,” says Lovell.
At first blush, a wider angle may be preferable, but Lovell argues that image resolution and fidelity to what is an officer's immediate vision is more important.
Although Vievu's cameras record in higher resolution, they do not offer any special features for recording at night. Taser's AXON body camera, in contrast, offers low light enhancement. Smith points out that many police incidents that would be of public interest tend to occur at night, making this a vital feature.
Lovell has a different take and argues that enhancing beyond a police officer's natural eyesight with the light available at the scene is effectively introducing new evidence into an encounter.
“We don't want the camera to see more than what the officer could have seen,” says Lovell. “Nor do we want the camera to see better at night. So we don't do any infrared or image stabliization. If we did then we'd be enhancing the image and that has implications for chain of custody and evidence standards. We'd be going beyond what the human eye can see and recording in formats that aren't original.”
Stanley says that low light image enhancement “would give additional evidence of what took place, but could confuse the issue of intention and knowledge on the part of officers,” which could then leave a jury with making the tough deliberation of whether an officer was actually able to see something that his camera recorded.
“The point is to capture what was seen by the people who were there,” says Stanley. “But on the other hand you can imagine scenarios in which a police officer could commit abuse in the dark.” He concludes that this is a new area of technology and policy and that these sorts of issues will have to be worked out in the years to come.
Perhaps the most contentious issue is what does and does not get recorded. Both models require officers manually to turn their cameras on and off as they respond to a call. (The oversized switch on Vievu's model looks to be slightly easier to manipulate, which could be useful when officers need to turn on a camera in a high-intensity situation.)
“A crucial and central point is that technology and policy needs to be designed so that police can't turn the video on and off at will,” says Stanley. “It should be remembered that even having video does not always constitute an objective record of events. It depends on when the video was turned on and when it was turned off.”
It's not hard to imagination a situation in which a police officer might intentionally shut off his camera and then later blame a supposed technical malfunction. Smith says that AXON cameras maintain logs on the device of when the cameras are manually turned on and off, which allows them to determine whether a shutoff was intentional or accidental.
Stanley says that one policy that has been suggested to deal with this problem “is to put the burden of proof on the police officer if there has been an accusation of abuse when the interaction should have been recorded but was not.”