Streaming to Protect and Serve: Video in Law Enforcement

These days, given the nature of streaming media and digital audio or video recordings, all that evidence is much easier to obtain and duplicate, but it's also much easier to lose track of or manipulate. There's a growing demand for indexing, search, and retrieval (ISR) solutions that not only maintain the chain of custody/chain of evidence (to eliminate the possibility a recording could be manipulated) but also make it easier to view or listen to media-based evidence.

As such, one area that's maturing alongside the growing stockpile of video evidence from body cameras and in-car recorders is the use of enhanced workflows that use metadata tagging systems to catalog video evidence while at the same time keeping it from being used in the wrong case.

Axon markets a system called Evidence.com that offers a cloud-based "evidence locker" for law enforcement departments overwhelmed by the capital expenditures required to maintain secure evidence servers to house thousands of hours of media for active cases. A key selling point of the Axon solution is a version of auto-tagging metadata against body cam or in-car recordings.

Axon's Evidence.com offers a cloud-based "evidence locker" for law enforcement agencies overwhelmed by the capital expenditures required to maintain secure evidence servers to house thousands of hours of media for active cases.

"We've observed that when busy officers manually tag videos with metadata, many videos are tagged with the incorrect information or aren't even tagged," Axon states in the frequently asked questions (FAQ) portion of its auto-tagging "how-to" documentation. 

One of Axon's key selling points is searchability of evidence.

"Video evidence can be invaluable, as long as it's easily logged and found," the company notes. And given the fact that interview room video also can be uploaded to the Evidence.com storage system, the ability to have all video content enables shift supervisors to manage evidence more effectively.

Even if the manual video tagging with key metadata is done correctly, there's an extensive effort required of officers that Axon argues is unecessary.

"Manually tagging a video takes up to 3 minutes of an officer's time," the company notes. "If officers record 5 videos per shift and work 16 shifts per month, that means each officer spends 4 hours per month entering metadata."

To solve this problem, Axon's auto-tagging subscription service will ingest a spreadsheet that's been exported from a law enforcement department's Computer Aided Dispatch (CAD) or Records Management System (RMS) and then correlate the metadata with videos already loaded in to Axon Evidence from body cameras the company sells to law enforcement agencies.

Given the highly sensitive nature of investigations, and the added burden "innocent until proven guilty" places on those who handle evidence—to keep it from leaking before the court systems take up the case—Axon does require a law enforcement agency to upload the spreadsheet of CAD and RMS data via an encryption application to the Axon Evidence.com servers.

Once the information reaches the Evidence.com servers, "customized software automatically ties the correct metadata to the appropriate videos" by cross-checking incident ID, category, location and other corresponding information.

One area where the data-metadata lines blur is in the area of transcripts used as evidence. To meet the growing requirement by some states' courts to provide a transcript of any audio recording (or video with an audio track) at the time that the recording is entered into evidence in a court case, a number of companies offer transcription services for law enforcement agencies.

A critical factor in handing off the recording to a third-party transcription service, though, is maintaining the chain of custody. This means that the simple act of uploading a video or audio recording into a third-party transcription portal must meet Criminal Justice Information Systems (CJIS) standards for encryption, security, and tracking of chain of custody requirements.

Even email presents an issue, as evidence sent through unencrypted email would not meet the CJIS minimum standard of 128-bit encryption and a 10-character-long decryption key.

In this way, CJIS-compliant cloud-based solutions offer a workflow alternative that tracks file-based evidence (including documents and audio or video recordings) through the process, including the creation of a transcript.

SpeakWrite.com services law enforcement agencies by providing fast turnaround on transcripts, typically within 24 hours, according to the company website.

SpeakWrite offers a transcription service specifically for law enforcement, using multifactor-authorization, authentication, and encryption to transport and store evidence, and transcripts can be ordered directly through Axon Evidence.com.

More important, though, at least for those law enforcement agencies that use a cloud-based evidence locker, the transcript can be ordered through the evidence portal and, upon completion of the transcript, automatically synced to the proper piece of evidence.

SpeakWrite requests that officers provide details such as the incident ID, number of suspects, and even names, to better enhance the likelihood of an accurate transcript—this is not yet a space for machine learning and other so-called "artificial intelligence," since many recordings occur in less-than-ideal acoustic settings—but even the toughest transcript requests on major holidays can be handled around the clock.

As part of the commitment to maintain CJIS compliance, SpeakWrite notes that it uses multi-factor authorization, authentication and encryption to transport and store evidence. In addition, those that access the evidence to provide transcriptions are part of "a network of fully vetted US – based typists, each of whom passed a background check, ensuring your evidence is secure."

Axon, which has partnered with SpeakWrite to offer a three-click ordering of transcriptions through Evidence.com also suggests that law enforcement agencies consider a secondary use for transcripts: summarizing an incident report.

"Officers can verbally summarize an incident using the Axon Capture mobile app, upload the audio, and then order a transcript of their audio summary that they can use to aid in their report writing," the company notes, adding that the easy-to-order transcription service is a "first step in reducing officer paperwork and enhancing evidence stored on Evidence.com."

Global Trends

When asked whether the trend in body cameras is just a United States and Canadian effort, Watchguard's Vanman says it is, for now.

"It is [primarily US and Canada] if you look at the dollars" said Vanman, "but there is adoption worldwide. We signed a deal last year with the Egyptian National Police, so we're in the midst of a rollout nationally in Egypt. We sold Iraq. We have deals in the UK and the Mexican Federale police, so we're starting to see significant traction internationally."

Areas for Improvement

The streaming media industry has an opportunity to help solve some of the problems that law enforcement faces with video-based evidence. There's a potential for streaming audio and video to be used in law enforcement efforts on a global basis, in regions of the world that may not subscribe to the American principle of "innocent before proven guilty," so it's up to those involved in streaming in the U.S. to get it right before these types of solutions are pushed on a global scale.

Without even delving into the use of streaming video in the court system—a topic we will cover in a future article, most likely in early 2020—three areas spring to mind where improvements could be beneficial. All three areas center on the chain of custody for video-based evidence.

The first is access to video evidence by the accused, so that there's an equal amount of time for suspects to view evidence that could appear truncated or contradictory, perhaps even before charges are filed. A simple case of this is traffic light cameras, which sometimes only show a fraction of the video content recorded during a green-to-yellow-to-red-light sequence. For criminal cases, this is even more critical, as truncated or manipulated videos (such as the growing trend for "deep fake" videos) means that significant time will be required for examining the video evidence for anomalies.

The second area is compatibility between evidence systems. This could manifest itself in simple ways, like the inability to ingest videos in proprietary formats used by some body camera manufacturers, but also in much bigger ways such as needed cross-agency collaboration. If evidence systems don't provide standardization of both ingest and export, the lowest common denominator version of a piece of video evidence will not only lose corresponding metadata but will also increase the number of versions of a single piece of video evidence that will need to be maintained on separate evidence system.

Finally, it seems the law enforcement industry is drowning in a sea of video-based evidence, none of which is inexpensive to retain. While there are appropriate tools in some evidence systems to automate standard operating procedure retention periods, there's a growing mountain of video evidence that will need to be maintained for cold cases and appeal cycles as cases wind their way through the court system over the span of a decade or more. We in the streaming industry know how to deliver and store video for equivalent lengths of time, so there's social justice and financial gain potential for those who assist law enforcement in their "long tail" problem.