Tag: Real

Long-Awaited Report Recommends Limits on Real Time Crime Center, Including Warrants for Investigators to Search Surveillance Footage

 

By Erica C. Barnett

The NYU Policing Project released its long-awaited privacy and civil-liberties audit of the Seattle Police Department’s Real Time Crime Center on Monday, one day before Mayor Katie Wilson plans to announce whether she will expand the city’s use of the surveillance cameras that feed into the center. The RTCC incorporates hundreds of city-operated and private surveillance cameras into a system that allows officers to surveil much of the city in real time, pulling in data from other sources to direct police in the field or investigate crimes after the fact.

The report did not make recommendations about whether the city should expand the center. Instead, the auditors made 19 recommendations to help address “shortcomings” in the program, including the potential for footage to wind up in the hands of federal agents and the lack of formal written policies governing things like data retention and when police can use the RTCC. These recommendations ranged from changes that would be fairly easy to implement—like ending SPD’s policy of retaining some camera footage indefinitely—to bigger lifts, like requiring a warrant any time police want to review someone’s movements retrospectively, which the auditors said constitutes a type of search.

“We think that there is a solid foundation for governance of the RTCC in Seattle… but that there are also significant gaps in policy that the city and SPD can fill,”  NYU Policing Project chief of staff Katie Kinsey said on Monday. “If the RTCC is going to continue to operate, it should do so subject to implementing these additional policy recommendations.”

Earlier this year, Mayor Katie Wilson “paused” the installation of new cameras on Capitol Hill and in the Central District; the city did install and turn cameras on near the stadiums just south of downtown, but only during the World Cup. Critics on both sides have assailed the mayor for her approach to cameras. Law-and-order officials like Councilmember Bob Kettle have harrumphed that she’s putting public safety at risk, while privacy advocates like Community Not Cameras have said police surveillance puts Seattle residents and visitors at risk.

Seattle’s RTCC, which began operations in 2025, provides police with real-time and after-the-fact access to footage from a network of  more than 500 individual cameras, many of them housed in single-location “units” with three to five cameras each, operated by SPD, the Seattle Department of Transportation, and private business owners who’ve volunteered to make their footage accessible to SPD. About 900 more private cameras, at 159 locations, are owned by people who’ve volunteered to make their footage accessible on request.

SPD officers at the RTCC access all these camera feeds using an Axon system called Fusus, which allows officers to call up camera footage and pull in other information to respond to ongoing incidents, overlaying that footage with law-enforcement databases, object-detection software, and other tools to give the police department a far more comprehensive picture of what’s happening in view of the cameras than old-fashioned CCTV cameras alone.

“When you’re connecting a bunch of cameras, when you’re having analysts that are able to look into them, when you’re running AI over them, we’re talking about really a different kind of technology,” Kinsey said.

SPD is barred from using facial recognition. But it does use a system called “object identification” that is supposed to be able to identify a list of object types, including vehicles, purses, and clothing colors. In practice, this system often fails to identify objects and people properly. In one high-profile case last year, police swarmed a 16-year-old student and handcuffed him after his school’s camera surveillance system mistook the bag of Doritos he was holding for a gun. (Speaking to the auditors about problems with the object identification system, Seattle police described a more innocuous example, in which a test search for a person wearing gray pants and a gray shirt turned up footage of a pigeon).

Most of the RTCC’s work involves supporting officers responding to 911 calls,  the report found, but officers working at the RTCC also use the system to investigate crimes by tracking people’s movements, going back through footage to surveil suspects “retrospectively.” The report recommends requiring police to obtain a warrant any time they want to use camera footage to investigate a crime after the fact, allowing “warrantless tracking” of individuals only when officers are responding to a situation in real time, to “prevent physical harm to the officers or other persons, the destruction of relevant evidence, or the escape of the suspect.”

The Policing Project is aware that this will be the most controversial recommendation, and took pains to point out that it would only apply to investigations that take place after a crime is over and the scene is cleared.

“We are not recommending a warrant in what are called exigent circumstances—when there is a public safety hazard, there’s a shooter, there’s someone who  represents a threat to public safety. Nor are we saying that SPD is powerless to check the cameras at the crime scene to see who may have been there,” said Max Isaacs, the Policing Project’s director of  technology law and policy.

Of all the recommendations, he said, the warrant requirement was “a hard pill for [SPD] to swallow. But “it is not the case that [a warrant requirement] would impede investigations. If the investigation is warranted, then they can always go to a judge and get signoff.”

The report also notes that many of SPD’s existing policies, like a commitment from Axon that it will disable new surveillance features by default, are “informal” and not a part of SPD’s policy manual or law, meaning that they can be changed at any time.

In one unwritten policy, SPD says it retains surveillance footage no longer than five days unless it’s needed for an investigation—but because there’s no formal policy, the department could increase that period to as long as 30 days without  any policy change. This casual approach is at odds with the seriousness of “‘indiscriminate’ surveillance,” the report found, and recommended that SPD adopt five-day retention as a formal policy. “We think [the lack of a written retention policy] is a mistake, and it should be made a matter of binding policy,” Isaacs said.

The auditors also recommended getting rid of footage used as evidence once it’s no longer needed. The current policy of keeping that footage indefinitely makes it more likely that bad actors, such as ICE or outside law enforcement, can file a public records request or issue a subpoena for SPD footage. “We think that creates unnecessary risk, and SPD should change it,” Isaacs said.

Other policies are vague to the point of meaninglessness—like an official policy that says the RTCC can be used to assist with “serious and violent” crimes. In practice, the report found, “SPD currently interprets” this phrase to include “misdemeanors of all kinds.” When there are no clear guardrails around how a police department can use surveillance, Kinsey said, “there is this tendency to have mission creep. So a tool that starts to be deployed or procured for one reason ends up being used for lots of other reasons because you have it on hand.” 

The report recommends setting clear standards for when police can access the RTCC, and says that in any future “review of the RTCC’s authorization, the City Council should consider whether to further limit the circumstances in which the RTCC [can] be activated.”

Police surveillance is a relatively new project in Seattle, unlike other cities where Flock-branded cameras have become a ubiquitous, often controversial feature of everyday life. (SPD’s cameras are from a different company, Axon, which also manufactures Tasers.) Proponents argue that the RTCC and its cameras deter crime and help solve crimes; in Seattle, they’ve been sold as solutions to sex trafficking, gun violence, drug sales, and more. But, as the NYU report notes, the evidence that they actually do any of these things is “thin.”

“SPD has published some data on RTCC impact, including that officers and detectives are three times more likely to arrest a suspect when they receive support from RTCC analysts.,” the report notes. “SPD statistics also reflect that in 2025, the RTCC assisted in 17 homicide cases, with its support helping to close 10 of these cases.”

“In SPD’s data analysis, however, it defines RTCC ‘support’ as a case in which an RTCC analyst takes any action (such as checking a camera), regardless of whether that action yielded any useful information or meaningfully impacted the investigation.” In other words, SPD is mixing up correlation and causation—they could just as easily claim that driving an SUV instead of a patrol car improved investigation outcomes, or that better in-car computers make it more likely that an officer will solve a crimde.

Wilson will make an announcement Tuesday morning about her plans for the RTCC. (On Monday, the mayor’s office declined to respond to written questions). For months, she’s been telegraphing that the report’s release is a final step before she announces plans to keep the RTCC and preserve or expand its surveillance footprint in Seattle. What’s unclear is which recommendations from the report she will try to put in place and whether she’ll make the RTCC’s ongoing existence and future expansion conditional on meeting them.

Speaking to PubliCola late Monday afternoon, Kinsey said she was “relatively optimistic about the adoption implementation of a lot of these recommendations, and for a few reasons. One is that we were asked to do this audit by the mayor’s office.  I think it always makes a really big difference when you’re doing something because leadership is actually asking you to do it, and not because some less empowered entity is asking you to do it.”

The other reason is that the Policing Project worked with stakeholders inside and outside the city for eight monts to come up with these recommendations. “Having a certain amount of buy-in, a certain amount of understanding through the process that these things are actually doable, helps me feel a little optimistic,” she said.

This is a developing story. We’ll have an update tomorrow after Wilson’s press conference, which will also feature interim police chief Andre Sayles.