As surveillance tools including drones and video cameras become more commonplace in public spaces nationwide, privacy advocates are calling on policymakers to impose greater scrutiny on the technologies. In Seattle, a 2017 surveillance ordinance that serves as a model for other U.S. cities seeking to govern such technologies is being put to the test.

But two years on, the ordinance has been bogged down by a bureaucratic process that has delayed City Council approval of technologies under review.

As the rollout of the ordinance gets under way, civil-liberties groups and an external working group that reviews the impacts of the technologies have raised concerns about the lack of clear restrictions on the collection, use or storage of data.

“There are a lot of eyes on Seattle as we build this airplane while we fly it,” said Shankar Narayan, a working-group member and ACLU of Washington’s technology and liberty project director. “And we believe for that reason it’s very important to get it right.”

The City Council committee that must approve each such technology before a full City Council vote met last month to discuss the first of the reports required by the ordinance. Those covered two surveillance technologies used to monitor and manage traffic by the Seattle Department of Transportation (SDOT): license-plate readers and closed-circuit television traffic cameras.

Council President Bruce Harrell and Councilmember Teresa Mosqueda, the two council members present at the meeting of the governance, equity and technology committee, put off a vote on the two SDOT technologies discussed, due in part to the dense 150 pages of each finalized report.


The session highlighted the need to streamline the surveillance impact reports to make the findings more palatable to the public and to make clearer the training guidelines for the operators of the technology, said Harrell.

Seattle’s look at surveillance technology comes at a time when jurisdictions throughout the nation are scrutinizing their own public-safety tools. Last month, San Francisco banned the use of facial-recognition technology used by all municipal agencies. Oakland, California and Somerville, Massachusetts are considering similar bans.

The 2017 Surveillance Ordinance seeks to offer greater transparency on the use of technologies that monitor or analyze people in a manner that could pose a risk to privacy or other civil liberties. It requires city departments to obtain council approval of surveillance technologies they now use, and those they plan to purchase.

City departments must write a surveillance impact report on each of the agency’s surveillance technologies. Each report identifies the policies around the use of a technology, the data that will be collected, and the staff who will operate the technology. A 2018 amendment of the law required that the impact report also be reviewed by an external working group consisting of seven members, most of whom represent marginalized communities historically surveilled by government agencies.

The current ordinance replaces a 2013 law restricting the acquisition of surveillance equipment. Public outrage following the city’s installation of dozens of security cameras along Seattle’s waterfront and a wireless mesh network by the police department spurred the creation of the original ordinance. In 2016, the Seattle Police Department purchased the social-media surveillance tool Geofeedia to aid in criminal investigations. The technology was not examined before acquisition since it was not hardware, underscoring the need to update the ordinance to include surveillance technology and data.

Since 2017, the city department now called Seattle Information Technology has gone through hundreds of technologies, whittling down the list of those requiring approval to 29 – the vast majority of which are used by the Seattle Police Department, followed equally by City Light, the Department of Transportation, and the Fire Department. Excluded from the law are police body cameras or technologies used in an office setting.


The creation of a surveillance impact report is a six- to seven-month process that consists of department staff reviewing the technology, writing a draft and organizing community outreach events to solicit feedback. An external working group then reviews the finalized report and responds with comments before the full City Council votes whether to approve the technology.

The technologies up for discussion included SDOT’s artificially intelligent license-plate readers that use high-speed cameras and computer algorithms to convert the images of license plates into data. Cameras positioned in 25 locations throughout the city capture a digital image of up to 10 percent of license plates that pass by. The state Department of Transportation matches the plates and recognizes when the same image appears at point A and point B, and uses that to generate the freeway displays informing drivers about the travel times on major corridors.

The Department of Transportation deletes the plate image after making the matches and SDOT never receives the information, said Seattle Department of Transportation representative Jason Cambridge. “Although these systems are similar to those used by law enforcement agencies, this is a completely separate system and it’s never used for that type of legal enforcement.”

The other SDOT technology up for City Council approval are the hundreds of remotely controlled traffic cameras designed to observe vehicle movement throughout the city. Both the cameras and license-plate readers have been essential in helping the public plan trips during a time of increased traffic and construction, said Cambridge.

Other technologies covered by the surveillance ordinance are more invasive. In April, the working group criticized the use of automatic license plate recognition technologies used by the Seattle Police Department that raised civil-liberty concerns. In an April letter to the Seattle City Council, the working group shared concerns about automated license plate reader systems used by the police department to collect 37,000 license plates in 24 hours, which adds up to 13.5 million scans a year.


The working group noted that the drivers scanned by the technology had not been suspected of any crimes, “which calls into question the scale and purpose of such data collection.” Seattle law doesn’t place limits on the use of automated license plate reader technology or data, meaning that any agency can choose how and when to track vehicles and retain data. The work group recommended that the Seattle Police Department clearly define its policies and cut the retention of automated license plate reader data to half of the current 90 days.

A letter to City Council committee members from Narayan and Rich Stolz, executive director of the immigrant and refugee advocacy nonprofit OneAmerica and a working-group member, said the reports contained outdated information and inaccuracies. The working group members urged the City Council not to adopt the lengthy reports as law. Ginger Armbruster, the city’s chief privacy officer, clarified that the department’s policies and practices of data collection, usage and retention were the only portions of the report that would be codified.

In an effort to make the reports digestible to the public, Mosqueda recommended that the departments create a three- to five-page summary document on policy practices to accompany their reports on surveillance technologies.

McKenna Lux, policy and advocacy coordinator of the Council on American-Islamic Relations-Washington, which is also represented in the working group, urged the City Council not to issue a blanket approval of SDOT’s use of license plate readers and CCTV traffic cameras.

“We’ve seen the consequences before of unregulated use of surveillance technologies, notably with the NYPD’s surveillance of the Muslim community in New York using tactics including license plate readers,” said Lux. “In order to prevent future abuse, we need to ensure that concerns raised by stakeholders and impacted communities are addressed in strong and comprehensive ways.”