Seattle’s New Policing Panopticon

A critical analysis from Puget Sound Prisoner Support looking at various surveillance technologies used by law enforcement in Seattle, WA.

On February 2nd, the City of Seattle quietly issued a batch of Surveillance Impact Reports (SIRs), required by the Surveillance Ordinance process prior to acquiring new surveillance technologies.

These reports signaled the City’s intent to acquire three separate technologies: Acoustic Gunshot Location Systems (AGLS), CCTV cameras, and Real-Time Crime Center (RTCC) software. While these reports are incredibly vague, those of us who have lived in Seattle for long enough can read between the lines.

The city claims these technologies will be targeting “gun violence, human trafficking, and persistent felony crime […] concentrated at specific geographic places”, beginning in the Chinatown-International District, Aurora, Downtown, and Belltown.

The City intends to combine these three separate technologies into a surveillance network that syncs CCTV footage with police location, 911 calls, and incident records. A closer look at these reports will show why we should all be concerned about the City’s impending acquisition of these technologies.

Acoustic Gunshot Location System (AGLS)

Anyone aware of Seattle Mayor Bruce Harrell knows that he has been a big proponent of ShotSpotter. Harrell has been pushing the technology for the past 10 years, at least as long as the company has been donating to his campaign funds.

Shotspotter is probably the most well-known of a class of systems the City is calling Acoustic Gunshot Location Systems (AGLS). As written in the SIR, AGLS use “microphones placed in a defined geographic area that are programmed to detect the sound of gunshots and alert police and 9-1-1 when and where the incident has taken place.” In other words, these systems dispatch police and EMS to locations of loud noises it classifies as gunshots.

These systems mostly use a combination of machine learning algorithms and human analysts to evaluate whether the loud noises its microphones pick up are actual gunfire or simply car backfires, fireworks, or other sounds. Training materials for ShotSpotter shows that its analysts have broad discretion in how they classify recordings.

Regarding privacy in public areas, the report claims that “audio recordings are only made if a sufficiently loud enough audio signal is detected (around 120 decibels)”. It continues, “[i]n the event that human conversation is inadvertently collected (for example, screaming that rises above 120 dB), SPD will work with the AGLS vendor to have those audio files and associated locations deleted from the system.”

One sentence in the report is key:

While the current literature on AGLS does not support its efficacy as a means to improve the speed and quality of police response, nor a means of enhanced reporting, no research to date has addressed the application of AGLS in the context in which SPD intends to deploy it as a component of a broader forensic tool to support criminal investigations. For that reason, the existing research on AGLS alone is not helpful.

To break this down, the City is preemptively admitting that it is aware of the many reports on AGLS showing that these technologies do not help police respond to gun-related incidents. Instead, it wants to use AGLS as a means to alert police of and collect forensic data on possible gunfire events.

Yet, past incidents already show how AGLS are (mis-)used by police as a “forensic tool.” Cases in both Chicago and Rochester show that alerts were reclassified as gunshots by analysts after ShotSpotter was contacted by police.

As an alerting system, ShotSpotter exacerbates police violence by sending police on high alert into our neighborhoods. A 2021 report by the Chicago Office of Inspector General showed that “evidence of a gun-related crime” was only found in “9.1% of CPD responses to ShotSpotter alerts”.

But for each of the other 90.9% of alerts, police are still dispatched. A video from 2023 showed a heavy SPD response to a 911 caller alleging shots fired, showing officers pointing their rifles toward an unarmed man in crisis. Officers finally left after a tense standoff with intervening community members. Every false positive is another possibly deadly encounter.

In Chicago, a ShotSpotter alert has already led to an officer shooting and killing 13-year-old Adam Toledo in 2021.

CCTV Cameras and Real-Time Crime Center (RTCC)

Just as concerning though, are the other two technologies the City is requesting: CCTV cameras and Real-Time Crime Center (RTCC) software.

The SIR proposes to not just install city-owned CCTV cameras, but also to allow “privately-owned security systems […] to voluntarily share video of storefronts and areas where the public has access” with the City. This would be a massive expansion of police surveillance, giving police potentially real-time surveillance video from any/every business concerned with petty theft or houseless people.

These feeds would then be integrated into Real-Time Crime Center software, which “provides a centralized location for real-time information and analysis” and “integrates dispatch, camera, officer location, gunshot detection, 911 calls, records management systems, and other information into one ‘pane of glass’ (a single view).”

Additionally, the RTCC software for CCTV cameras “can also provide in-application video analytics that use machine-learned algorithms to analyze camera feeds and, using object recognition, locate specific items, people based on clothing, or vehicles based on description”.

Consider Fusus, the RTCC software that most closely matches the capabilities laid out in the SIR. (Other RTCC products by Genetec and Flock support most of the same basic features)

The Fusus camera registry allows business owners and community members to grant varying levels of access to privately-owned cameras, including Ring doorbells. Depending on the level of access granted, camera feeds can be “shared all the time, automatically activated by a trigger, or only on emergency alert.” So in many cases, instead of having to get a warrant or ask camera owners to send over video files, police will have perpetual live access to video from participating cameras.

Training materials from the Atlanta Police Department show the ability to search camera footage for vehicles (by type, make, color, etc), people (by gender, clothing color, whether the person is wearing a hat or bag, etc). Keep in mind that these criteria will only continue to develop as time goes on. Users can also choose to be notified whenever any person or vehicle matching the criteria are detected in the search area.

Fusus also has the ability to turn every camera into a license plate reader. This would continue the massive ALPR expansion that was proposed in December. Along with the alarming amount of information that can be inferred from aggregating a person’s travel patterns, experts worry about the dangers ALPRs pose to immigrants and people seeking reproductive care.

The software can also integrate with an ever-expanding list of third-party police technologies such as ShotSpotter to stockpile even more data.

Aside from the enormous incursions on privacy these technologies would bring, you would be hard-pressed find a worse group of people to give this power to than the Seattle Police Department. In just the past five years, one officer was fired for using internal police data to try to get a date, while another was fired for using police databases to stalk his ex-girlfriend. This isn’t to mention the multiple domestic violence investigations into officers in the department.

Conclusion

Taken together, these technologies paint a picture, not of a safer city, but of a compassion-less city dictated by economics under the eye of unblinking cameras.

We oppose any and all forms of police surveillance, urge those living in Seattle to question their desires for “public safety” and interrogate what that actually means, and to stand in solidarity with our marginalized neighbors who will surely be the targets of this new form of police terror.

Photo by Milan Malkomes on Unsplash