Police Body Worn Cameras – they are only effective if they are actually recording!

Standard

The shooting this week of a 40-year-old Australian woman, Justine Damond, by a Minneapolis Police Department officer has highlighted once again the shortcomings with Body Worn Cameras (BWC). In a case where everyone is searching for answers as to why a woman in her pyjamas was shot dead by a police officer, the body worn cameras, issued to both the officers in the vehicle and which should have provided the answers, were not active.

Unfortunately this is not the first such incident where police officers have been involved in a fatal shooting and the cameras have not been active or failed to provide evidence of a suitable quality.

  • In 2016 Alton Stirling was fatally shot by Baton Rouge police. Both officers involved in the event were equipped with BWCs but it is widely reported that officials claim the cameras fell off during the struggle and did not record the incident, although bystanders and CCTV did capture footage.

 

  • Keith Lamont Scott was fatally shot in September 2016 by a Charlotte-Mecklenburg Police Department Office. In this case, the officer who shot Scott was not wearing a camera and another officer present and who was wearing a camera did not activate it until after the shot was fired, reports the Washington Post. Charlotte-Mecklenburg Police Department published the BWC footage from the incident (below); note the beep on the audio which indicates when the recording was activated.

  • On March 14th 2017, a Freemont detective fatally shot a 16-year-old Hayward girl during an undercover operation. The detective, who normally wears a body cam, didn’t do so on that day. “My preliminary understanding is that it wasn’t working and so they weren’t using it that day,” Freemont Chief of Police Richard Lucero said in a recent interview with The Argus, referring to the March 14 shooting that ended with Antioch teenager Elena Mondragon dead.

These types of incident are controversial because, particularly with cases such as that of Justine Damond, the circumstances raise a suspicion in the minds of the public that there was a deliberate intent to not record the events taking place. Whilst the official investigation will hopefully shed further light on this and provide clarity, the way this technology is currently used will always hold the potential for events not to be recorded as it relies on manual activation by the wearer.  As a result and due to factors such as the high stress situation of the rapidly unfolding events, lack of training or deliberate intent, recording may not be started when it should be.

Why do Police wear body cameras?

Although the policies relating to the use of body worn cameras may vary between different police forces/departments, the motivation for their use is broadly the same: to record interactions with members of the public that have a criminal or investigative purpose.  Therefore, they are not intended to record everything the officer does nor the interactions with the public that do not fit this purpose e.g. social interaction (saying hello), giving directions etc.

Nonetheless, the use of these cameras should be beneficial to both the police officers wearing them and the members of the public. They are intended to help the police by providing a clear record of any criminal activity that they witness or are involved in investigating, thereby providing valuable evidential data and a clear and indisputable record of the events that took place. For the members of the public, they should provide increased confidence that any interaction that have with police is conducted professionally, within the scope of the law and the recordings will support their case if they consider the interaction is unreasonable or unlawful.

When should the Body Worn Camera record?

Body worn cameras are not normally in operation all the time and recording has to be manually activated by the wearer. Although there are a number of different brands of camera available, they all typically work in the same way. At the start of shift the officer will retrieve a camera and install it in the mounting on their uniform.  It is then turned on which places the camera into standby mode. In order to start recording, the officer must press once or twice or a button on the device; a clear audio and visual indication that the device is recording is typically provided to alert both the wearer and the public nearby.

Some cameras, such as those made by Axon, and commonly used by police forces around the world, continuously record in the standby mode and as soon as the camera is placed in record mode, the preceding 30 seconds of video (but not audio) are also saved.

The occasions on which recording should be activated will be the subject of each individual police force’s policy. For the Charlotte-Mecklenburg Police Department (Keith Scott incident) their policy states that officers must fully activate their body cameras “prior to or in anticipation of” interactions with civilians resulting from traffic stops, suspicious vehicle or persons investigations, arrests, use of force incidents and voluntary investigative conduct. In the policy, “voluntary investigative conduct” is defined as “the mere suspicion of some type of criminal activity by a person,” which fits the description of why police say officers confronted Scott.

Why are the cameras not recording all the time?

It is not unreasonable to ask the question why BWCs are not recording all the time that an officer is on shift. In fact, there are a number of reasons why the cameras do not record all the time. These reasons include:

  • Police policy (see above) and privacy laws that place limitations on when the cameras should be recording.
  • The police officers themselves, even whilst on duty have a right to a reasonable amount of privacy. An obvious example of this is when using the bathroom. However, when not actively responding to crimes the police officers must be able to interact amongst themselves and discuss, for example, matters of internal policy without fear on these conversations being monitored with potential reprisals from their superiors.
  • Given than recording from the cameras are stored, usually for at least 30 days if not needed for evidence, the storage requirements would be enormous and add substantially to the cost of the system if the cameras were operating continuously.  As it is, the costs of storing the current volume of video is already very significant for police forces, often running into millions of dollars depending on the number of units in use.
  • Investigations would take longer as finding the actual events of interest would become more complex.
  • Civil liberties and perception issues; recording continuously will only further increase the public perception of the big-brother state. Whilst the use of video surveillance is becoming increasingly prevalent, this must be balanced as reasonably as possible with the right to privacy.

Why might recording fail to be activated and what can be done about this?

Whilst it is not clear exactly how often police officers fail to activate recording in contravention of policy, since typically the public only become aware of high profile incidents such as fatal shootings, the reasons why recoding is not started may be numerous. Some reasons may be (not including technical failures):

  • The officer simply forgot. This may be because they don’t have a need to regularly start recording due to a low level of incidents and therefore they are not in the habit.
  • The officer has insufficient training, including scenario based trained, to build up a reflex response to pressing the buttons to start recording.
  • There may be a conflict with the officer’s instinct to first seize a weapon to defend themselves in a potentially violent situation.
  • The officer may deliberately not want their activities recorded;  this is the most controversial point as it implies the officer is expecting to have an unethical or unlawful interaction.

Police forces/departments should ensure that a clear policy concerning the use of BWCs is published to all concerned personnel and that each contravention of the policy is taken seriously, logged, investigated and addressed with the concerned officer. If it is found that reasons such as the first three above are contributing to a significant number of encounters not being recorded, this can largely be addressed through further training and/or modification of procedures.  It may also be necessary to review the location where the camera is mounted to ensure the best coverage as well as easy access to the recording button.

Some police forces have already implemented procedures that require the dispatcher to actively confirm with the officers that their BWCs are recording. This is probably something that should be implemented within the standard operating procedures of all forces using BWCs.

Finally, persistent and/or deliberate non-compliance must be made a formal disciplinary matter.

Why can’t the activation of recording be automated?

In order to automate the start of recording it is necessary to have both a trigger to determine that recording should start and a means for the camera to be aware of this trigger or receive an instruction resulting from the trigger to start recording. It should be born in mind, however, that most BWCs are currently very basic and don’t have the means to communicate with external triggers. Therefore, any practical solution will require time to develop and implement in a next generation camera along with any supporting infrastructure needed. These in turn will increase the capital and on-going operational costs of the solution.

In addition, any automation should not replace the officer being expected to start the camera recording themselves whenever possible, but should serve as a supplementary feature to assure that recording starts.

Below we discuss some of the possible triggers and what would be needed to allow this trigger to start the camera recording.

A signal from the dispatching systems

Police officers are typically responding to an incident that is notified by a dispatcher, or they notify the dispatcher that they are attending an incident that they have become aware of.   Therefore, since the dispatcher will usually be aware of which officers are attending an incident it should be possible for them to send automatically, via their incident/resource management systems, a signal to instruct the camera to start recording. This can be used in conjunction with procedures that say that the officer must attempt to manually start recording.

Technical implications & limitations

The current generations of BWCs currently usually don’t have any form of wide-area network connectivity. Therefore, they will need some form of connectivity back to the dispatching systems. Adding 4G/LTE connectivity would provide this. Alternatively it may be possible to integrate the camera to the officer’s radio using Bluetooth for a low bandwidth connection.

Pros of this approach:

  • With 4G/LTE this would provide a high bandwidth connection that would also allow the device to be accessed remotely for real time streaming or downloading of video whilst the officer is still on deployment.
  • Will allow the status and health of the device to be communicated therefore the dispatchers will be able to see if an office turns off recording during an incident or if the unit has developed a fault, the battery is becoming depleted etc.

Cons of this approach:

  • It will add to the capital (purchase) cost (possible substantially) of each unit.
  • For 4G/LTE, it will require a data plan on the device from a third party provider if the force does not have their own suitable network.  This could be a large cost item.
  • For Bluetooth, the integration may be complex and radio vendor dependent. In addition there may be complex logistical issues as the radio will need to paired to the camera.
  • 4G/LTE will be dependent on coverage so may not work reliably in certain locations such as in buildings.
  • May significantly increase the power consumption and thereby reduce the battery life of each device.

A signal from the vehicle for vehicle based officers

For officers who are vehicle based it should be possible to generate a trigger from vehicle events such as moving at high speed or when the light bar is activated. In fact, Axon already market a product called Axon Signal which allows a range of their cameras to be automatically activated with vehicle triggers (and other triggers such as the use of their Taser device).

Technical implications & limitations

This solution will require a unit to be installed in the vehicle and connected to any relevant and reliable triggers.

Pros

  • Should provide reliable activation in many common cases for vehicle-based officers.
  • A likely lower cost of implementation than with network enabled devices.

Cons

  • May not work if the officer has left the vehicle and is out of range.
  • Will only for vehicle triggers. Therefore, if the officer has left the vehicle and then events unfold what will require recording, this solution is less likely to be of benefit.
  • Will be of less relevant to forces or divines that have a majority of personnel who are not vehicle based.

Use of accelerometers and other sensors to identify an officer’s reaction to an event.

When a police officer responds to an event it if usually accompanied by a sudden change in activity. For example, they will start running or making non-typical movements. In addition it is likely that there will be a change in their body such as an increased heart rate.  Furthermore, they may automatically draw a baton or other weapon that would indicate a conflict situation.

Technical implications & limitations

Detecting changes in body characteristics would require an additional sensor to be worn, and detecting the drawing of a weapon would require a sensor to be incorporated into the holster.  On the other hand, an accelerometer incorporated into the camera itself should be able to detect unusual movement or running and therefore automatically trigger recording. Indeed, Bodyworn is one manufacturer who is already claiming to do this with their device.

Pros

  • Accelerometers should be relatively low cost to implement in devices.
  • Does not require an external communication link (unless used with a body or holster sensor)
  • Should be software modifiable allowing for algorithms to be adapted and improved over time.
  • Can be expected to be reliable in detecting abnormal reactions to events.
  • Should be able to produce a trigger such as an officer falling down or staggering after an unprovoked attack (combined with continuous recording of the 30 seconds or so leading up to this trigger could produce a very clear picture of what just happened even when the officer is unable to react in time).

Cons

  • May not work reliably or need different profiles for officers based on horses, bicycles etc.
  • Holster triggers, in particular may result in more equipment that the officer has to wear, would have to be extremely durable but most likely will be prone to frequent failure.

The impact of Body Worn Cameras & Public Perception

It is interesting to see that the American Civil Liberties Union (ACLU), which is generally opposed to ‘pervasive government surveillance’ has come out broadly in support of the police use of body worn cameras, “but only if they are deployed within a framework of strong policies to ensure they protect the public without becoming yet another system for routine surveillance of the public”. The ACLU recognises that BWCs can be a win-win helping to protect the public against police misconduct, and at the same time protect the police against false claims of abuse.

A study by Cambridge University’s Institute of Criminology, shows that the use by officers of body-worn cameras is associated with a startling 93% reduction in citizen complaints against police. Researchers say that this may be down to the cameras modifying the behaviour through an ‘observer effect’: the awareness that encounters are recorded improves both suspect demeanour and police procedural compliance. Essentially, the “digital witness” of the camera encourages cooler heads to prevail.

The findings from Cambridge University are further supported by those from a study by Florida Atlantic University. Key findings from this study reveal that 87 per cent of respondents agree that BWCs would improve police officer behaviour and that 70 per cent agree that BWCs would improve how citizens behave when they en‐ counter police. What was a but surprising from this study was their expectation that people with the most negative views of the police would be most supportive of BWCs. Surprisingly, they found the opposite to be true. Citizens who had a more positive view of police and thought they were treating people fairly and doing a good job had the most support for BWCs.

Another unexpected result of the study was that those citizens who were the most concerned about crime were less inclined to see benefits in the use of BWCs. However, the researchers caution that this is an indirect relationship having to do with their perceptions of police performance, fear of crime, and belief that police are not doing a good job and therefore they perceive less benefits of using BWCs.

In addition to benefit of increased transparency in the police’s interactions with the public, body worn cameras can also be used to positively show the role of the police and the situations officers find themselves in. According to a report in the Christian Science Monitor, in March 2015, four Cleveland Police Department officers were involved in a shooting incident that resulted in patrolman David Muniz being shot as they climbed the stairs to Theodore Johnson’s apartment. As Office Muniz turns a corner, Mr. Johnson opened fire and the office is recorded on video saying “I’ve been hit”. Further recordings from Muniz’s cam show him and other officers pleading with Johnson to put the gun down whilst Johnson is pleading with the officers to kill him.

Another officer can be heard telling Johnson to “put the gun down and we’ll get you all the help you need.”

“I know you shot me, but I’m not going to shoot you,” says Muniz, who is standing just a few feet from Johnson.

Johnson still refuses to drop the gun, which he is holding at his side.

A few seconds later he raises the gun, and the officers open fire.

A grand jury determined that the four patrolmen were justified in using lethal force against Johnson. They had gone to Johnson’s home that night after his wife went to a police station and reported that Johnson had threatened to kill both her and their landlady.

Costs of deployment

The deployment of body worn cameras by police forces is not insignificant, with just the units themselves costing of the order US$500 for a standard model. On top of this there are the accessories, the docking equipment and the video storage and evidence management software that in some cases, such as Axon’s Evidence.com, requires a subscription.  Annual support and maintenance costs will also have to be factored as well as spares for loss, training, force expansion and irreparable damage.  A few examples of some recent BWC project scope and costs include:

  • In November 2015, the London Mayor’s Office for Policing and Crime (MOPAC), following a successful trial awarded a three-year contract worth £3.4 million to Axon Public Safety UK Limited, to supply the MPS with 22,000 cameras.
  • On March 7, 2017, the Sacramento City Council passed a resolution suspending competitive bidding in the best interests of the City for the purchase of a body worn cameras and digital media storage solution and authorizing the City Manager or City Manager’s designee to enter into a contract with TASER International, Inc.  The council report identifies the terms of the agreement include a five-year time of performance for a not-to-exceed amount of $3,956,628 with initial deployment phases of $578,501 and $912,695 during FY2016/17 and FY2017/18 and subsequent annual costs of $821,811 from FY2018/19 through FY2020/21.  The council report indicates that this is for the supply of 750 cameras with 50 additional reserve cameras and a further 70 training cameras.
  • A report in the register states that Greater Manchester Police are investing over £2m over 3 years for 3000 BWCs and the associated storage and support.

Body Worn Cameras are big business for the vendors and there is no doubt that the demand for them is only going to increase.  Nonetheless, governments and police authorities have to address where the funding is coming from, with some police department having to make a choice on whether to spend their limited funds on police officers on the street or storage technology. With storage costs alone often running into millions of US$, this can be a tough decision.

The choice of technology and system is also crucial in making sure of a successful project that delivers the most value. In 2015 The Texarkana Texas Police Department suspended their BWC project due to the large number of technical issues and camera failures. This resulted in them returning all the cameras to the manufacturer L3-Vision for a refund.

Future trends

Given the benefits already identified to the public and police of BWCs, this is a market that is going to see many increases in functionality in the future.   It is only a matter of time before live streaming becomes part and parcel of the BWC solution, although careful consideration of the functionality and performance of this solution needs to be made against it’s likely step increase in cost.

Other functionality such as live facial recognition is starting to become available with digital barriers announcing this month that their SmartVis facial recognition software has been adapted to run on the company’s body worn cameras.  This type of technology, if well implemented, will help the police improve their performance by identifying suspects that would have otherwise gone undetected.  However, wide spread use of this of technology can be expected to meet with strong public opposition if strong safeguards are not put in place from the beginning to ensure that citizen’s rights to privacy are protected.

Download a pdf version of this article here