Regional law enforcement agencies have used facial recognition technology for years to help them identify people who are either unwilling or unable to identify themselves. To do that, they run images taken with mobile devices against a database of nearly two million mugshots, looking for a match.
A new state law that goes into effect in the new year will put a temporary end to the practice, which some say poses ethical, security and privacy risks.
Opinions differ among law enforcement authorities in San Diego County as to how the new law will impact police work, but most agree that they are losing an efficient and effective tool. As a result, they’ll have to switch to more time-consuming methods of identifying people, like fingerprinting.
And that, some say, could have an effect on public safety.
“Here was technology working at a quicker pace for us to make sure we identified people correctly… (to) get our officers back on the streets quicker to patrol and keep our community safe,” Chula Vista police Chief Roxana Kennedy said in an interview this week.
Others say the change in the law, while disappointing to some, isn’t likely to cause big problems.
“Facial recognition is still developing, so it hasn’t gotten to the point where law enforcement is fully dependent on it,” said Lt. Justin White, a spokesman for the San Diego County Sheriff’s Department. “We still have traditional ways of doing police work and identifying people. Facial recognition obviously makes things faster and can narrow down the scope faster than traditional ways, but it’s not the end-all of identifying people.”
Assembly Bill 1215, prohibits officers and deputies throughout the state from using cameras for facial recognition and biometric surveillance. Gov. Gavin Newsom signed the bill into law in October. It will take effect Jan. 1.
AB 1215 comes as use of the technology has increased around the United States. Federal agencies use facial recognition at a variety of places including international ports of entry and airports, sparking concern among some segments of the public about privacy and civil rights issues. San Francisco and Oakland were among the first cities in the country to place specific bans on facial recognition technology, as was Somerville, Massachusetts.
In May, San Francisco passed its “Stop Secret Surveillance” ordinance, which banned city government agencies from using facial recognition technology. Months later, the Oakland City Council passed a similar ordinance, which prohibits city agencies, including police, from “acquiring, obtaining, retaining, requesting, or accessing” facial recognition technology, including if used by other police agencies.
In San Diego County, the three-year moratorium imposed by AB 1215 affects a large regional database known as the Tactical Identification System, or TACIDS, access to which is overseen by the San Diego Association of Governments. It contains roughly 1.8 million booking photos and is used by about 30 agencies in the region, including Immigration and Customs Enforcement and the Border Patrol.
About two weeks after Newsom signed the bill, the digital and civil liberties watchdog group Electronic Frontier Foundation sent a letter to SANDAG, asking the agency to “begin the process immediately to suspend this fatally flawed program that threatens the civil liberties of people in California”
“This is huge news that they are shutting it down,” said Dave Maass, a senior investigator with the EFF and author of the letter. “San Diego has one of the longest-running face recognition programs, it’s one of the most expansive with 25,000 scans a year. While we have seen bans on and moratoriums on facial recognition, it may be the only time a face recognition program of this size and scale was so abruptly shut down.”
Facial recognition technology is a part of a larger, ongoing conversation about biometric data — which refers to physical or behavioral information, including fingerprints, retinal scans and DNA — and how it may be collected, stored and used in surveillance.
The FBI says its database — called the Next Generation Identification system — is “the world’s largest and most efficient electronic repository of biometric and criminal history information,” and is used to help the agency and its partners combat crime and terrorism, according to its website.
But organizations like the The American Civil Liberties Union see red flags. The ACLU describes government agencies’ accumulation of such types of information as “an extraordinary threat to privacy.” The goal of the FBI’s biometrics database, the ACLU says on its website, is “to allow every police officer in the country to be able to instantaneously identify any individual in their crosshairs.”
The language of AB 1215 states that using biometric surveillance violates constitutional rights because it is the “functional equivalent” of requiring people to carry identification at all times. California does not have the “Stop and Identify” law, meaning if an officer asks for a person’s ID, that person cannot be charged for declining to provide it.
Some critics argue that simply storing such information in a searchable database makes it hard to keep out of the wrong hands. In June, the U.S. Customs and Border Protection suffered a “malicious cyberattack” that left thousands of photos of travelers and license plates exposed. A California DMV data breach allowed federal agencies, including the U.S. Department of Homeland Security, improper access to 3,200 Social Security numbers last month.
The technology itself is imperfect. While facial recognition has been useful in authenticating or helping to figure out a person’s identity, it has proven less effective when identifying women and people of color. The Washington Post reported that Asian people and black people are up to 100 times more likely to be misidentified by the technology than white men. Black women were more often falsely identified in searches by police that could lead to the women being arrested and falsely charged for a crime. According to the study the Post cited, the accuracy rates of the technology varied widely.
TACIDS is a part of a larger database called the Automated Regional Justice Information System, or ARJIS. The latter contains statistics and other crime-solving tools used by local, state and federal agencies within the region. Use of TACIDS has increased dramatically since 2013, when local law enforcement accessed the database approximately 6,000 times. Now, local agencies use it four times as often, roughly 25,000 times in 2017 and 2018 respectively, according to SANDAG.
The departments that used it most during both those years were:
San Diego Police Department: 243 officers used it 14,050 times.La Mesa Police Department: 49 officers used it 5,757 times.San Diego Harbor Police: 112 officers used it 5,049 times.Chula Vista Police Department: 140 officers used it 4,937 times.El Cajon Police Department: 68 officers used it 3,606 times.Kennedy, the Chula Vista police chief, said she understands the concerns about privacy, but noted that there is a prevalent misconception among members of the public that local law enforcement is using the technology to constantly watch people.
“There is no random surveillance,” she said.
Lt. Shawn Takeuchi of the San Diego Police Department said the moratorium is “unfortunate” but he could not predict how losing TACIDS will affect the department. Although SDPD was the most frequent user of the database, Takeuchi said, TACIDS was limiting because it contained only mugshots from San Diego County. The fingerprint database has a nationwide scope.
SANDAG’s Public Safety Committee is expected to discuss AB 1215 at its Friday meeting. A spokeswoman for the agency said it would work with law enforcement to ensure they are compliant with the new law.
© 2019 The San Diego Union-Tribune
Distributed by Tribune Content Agency, LLC.