13 August 2015

Televising the Revolution

August 3, 2015

Some revolutions change humanity for better, some for the worse, but they all share this: You don't know they've started when you're in them. On August 1, 1981, Buggles' "Video Killed the Radio Star" was the debut music video on the new Music Television Network, MTV. It was probably seen by less than 1,000 people that very day, but ask anyone under 50 now about music without video and you'll get a blank stare. 

Likewise, for good or ill, police Body Worn Cameras, or BWCs, are here to stay. In the United States, racially tinged incidents of police-public interaction -- from Ferguson, to Baltimore, to South Carolina -- have generated a broad, bipartisan consensus that all police officers should wear BWCs. The Scott shooting video from South Carolina (recorded not on a BWC, but on a bystander's phone) suggests that BWCs could deter at least the worst police behavior. As the song says, pictures came and broke our hearts.

In the United Kingdom, we are already more accustomed to video surveillance than are people in the United States. But two incidents have increased the momentum for the wide deployment of BWCs: the 2003 killing, at the hands of police, of Charles de Menenzes, which is now being compared to the Walter Scott case by the European Court of Human Rights; and the plainly less tragic but more-discussed "plebgate" incident. BWCs can protect us from those meant to protect us.

Still, consensus does not equal risk-free virtue. Ubiquitous BWCs can enhance the administration of justice and the deterrence of bad behavior, by police and by civilians. But they also risk serious erosion of our privacy, civil liberties, and human rights, and, indeed, they put at risk our operation as free societies. Law enforcement agencies and their political masters must develop and implement policies, procedures, and technologies to maximize the benefits of BWCs and to mitigate the above-mentioned risks. 

This needs to happen now, not after thousands of terabytes of video and audio are recorded, stored, processed, and used. By one estimate, a single American police officer recording all citizen contacts during a shift will generate 1.5 terabytes of data a year. Run the numbers for any decent-sized police force and you'll soon be into petabytes of BWC data, much of it recording the most private, intimate, and embarrassing moments we can experience: the aftermath of rape, domestic violence incidents, serious health crises and traumatic injuries to name a few. 

So, first principles for any BWC deployment:

Silence is Golden? Many police agencies dipping a toe into BWCs are, almost by default, opting to enable audio recording along with video. We understand the rationale behind such decisions, but constant police audio recording dramatically escalates privacy and civil liberties risks, not only to criminal suspects and others directly involved in interactions with police, but also to the myriad other, innocent people the police encounter every day. To the extent audio is enabled, mechanisms to obtain the consent of those being recorded and to protect suspects' right to remain silent, where applicable, must be implemented. 

Who's in Charge? Police not only have the power to deprive citizens of their liberty, they routinely observe some of the most intimate, personal, and embarrassing moments of our lives. From this perspective, enabling individual officers, or police department policymakers, to control when BWCs are turned on and off is a good thing. If individual officers are given discretion to do so, however, there likely will be endless litigation over whether, in specific cases, officers turned the BWC on or off to further their own interests, conceal evidence of malfeasance, etc., or for legitimate privacy reasons. We need rules to determine how this will be handled in practice. 

It's Not Just About the Hardware. Cost, reliability, control, and other issues all affect the purchase and deployment of cameras and of related equipment. But important privacy and civil liberties risks arise not primarily from the hardware itself, but from key policy issues that must be resolved at the beginning of any BWC deployment, including: 

Access & Disclosure. Access to BWC data should be limited only to those with a legitimate need for it, e.g., police, prosecutors, defense attorneys (and their clients), and judges. Access to BWC data for unrelated civil litigation or other purposes, however legitimate, should be extremely limited. One issue that has driven a great deal of public comment and disagreement in the United States is whether BWC data should be subject to "sunshine laws" or other mechanisms for journalists and the public to gain access to government-held data. Though mindful of the importance of open government, on balance, issues of individual privacy must outweigh such concerns, and BWC data should be subject to blanket exemptions from public disclosure. The possible exception is in extraordinary cases of public interest, and then only when individual privacy is protected. 

Retention and Destruction. BWC recordings should be retained only as long as is reasonably necessary to bring, and finally resolve, criminal proceedings, and to resolve allegations of misconduct. Such recordings should be destroyed automatically after a reasonable time, unless an accountable official directs their retention, and then only for narrowly circumscribed purposes. 

Analysis and Repurposing. Analysis of stored BWC data should be powerful, reliable, and utilize strong privacy and civil liberty protections. This means the analysis is fully audited, and is permitted only to resolve criminal cases and police misconduct claims. Strong security and audit controls should deter, prevent, and, if the first two fail, mitigate and punish repurposing or unauthorized sharing of BWC data. 

Consistent Standards. To the greatest extent possible, storage, analytics, use, retention, destruction, security, and related issues should be consistent across police agencies and jurisdictions, and should, at a minimum, comply with relevant, pre-existing standards and best practices, e.g. the FBI's Criminal Justice Information Systems requirements which, among other things, provides requirements for US law enforcement agencies handling of criminal justice information, including in the cloud. 

No Commercial Data Mining. To the extent that police store BWC data outside of their own control, particularly with private cloud-service providers, legal and technical measures must prevent private companies from mining or otherwise using BWC data for commercial or other unauthorized purposes. 

Broad deployment of BWCs will not be a panacea; nor will individual resultant recordings, in any particular case, depict the "whole truth." But better technology and more data in pursuit of finding the truth has historically led to greater justice. We believe this can happen with BWCs, if, but only if, those in power make hard decisions on at least the issues highlighted above, and sooner rather than later. If leaders fail to do so at the front end of the BWC revolution, the rights of many will be sacrificed.

Either way, this revolution will be televised.

No comments: