“Facial recognition technology in use,” it would say, under legislation proposed in Baltimore City Council.
The signs would warn that entering the establishment amounts to “consenting to your physical appearance being collected, generated or analyzed.”
City law enforcement fall under the proposal, too.
Police would be able to use facial recognition tools only to investigate the most serious crimes. The agencies — Baltimore Police, along with police forces for private or academic institutions — would be required to document the frequency of the tools’ use in annual reports. And the specific technologies used would have to undergo certain evaluations for accuracy.
Two Baltimore City Council committees on Wednesday heard discussion about a pair of proposals designed to regulate the growing technology within city boundaries. Neither held a vote, and their sponsor, Councilman Kristerfer Burnett, said there’s further tweaks are needed to finalize the language.
Still, Burnett emphasized the issue’s importance, calling the proposals a “step in the right direction” in an interview following the hearings.
“The status quo is there are no legal guardrails in the private sector or public sector on how most of these technologies are used,” except for existing state rules on public information and data sharing, Burnett said. “Beyond that, there’s really nothing in place that protects civil liberties, civil rights.”
Facial recognition systems use artificial intelligence to generate a unique identifier for an individual’s face, based on key features such as their ear shape or nose width. When fed a photo, the image is compared to previously stored data and a series of possible matches are generated. A person then analyzes the options for a match.
Critics say the technology, unregulated, can lead to privacy concerns, misidentification and lack of transparency. Additionally, the systems have been shown to be less accurate identifying Black and brown faces, potentially exacerbating harm for those populations, which already face discrimination and disparities throughout the criminal justice system.
In one recent criminal case out of Baltimore County, discussed Wednesday, a man who was misidentified by an analyst using facial recognition tools spent days in jail. The case, first reported by Wired and the New Yorker, was confirmed by Baltimore County State’s Attorney Scott Shellenberger.
Shellenberger said it was his understanding that facial recognition technology suggested the man’s identity, and authorities showed his picture to his probation officer. The man’s wife “worked very hard to convince police that it wasn’t him,” Shellenberger said. After officials confirmed it wasn’t him, prosecutors were notified, and he was released from jail the next day, he added.
Burnett’s proposals, which were scaled back from the more restrictive bill he filed this spring following input from the city’s law department, would foster greater transparency and understanding about the use of such tools in the city, he and supporters said Wednesday.
His second piece of legislation would establish an 11-member community advisory commission on surveillance technology, to provide policy recommendations and oversight, and establish “extensive” reporting requirements for city agencies.
That reporting would include facial recognition tools, along with other surveillance technologies they deploy, such as cell site simulators known as “stingrays” that imitate a cell phone tower and allow law enforcement to gather information; license plate readers; CCTV cameras; gun detection software; and body cameras, Burnett said Wednesday.
“This pulls back the veil of the utilization of surveillance technology, so that the citizens have a better understanding of what in the city is being done, but also have a voice in the procurement process, even without an actual vote,” Burnett said.
Andrew Northrup, from the Maryland Office of the Public Defender, told city council members that cases in which law enforcement used surveillance technology have “risen exponentially” during his time with the office, and that the tools are often undisclosed or described in vague language.
For instance, Northrup said, Baltimore Police received about 811 facial recognition reports in 2022, according to information the office obtained through a public records request. But public defenders were aware of only “a fraction of these,” he said, calling that “alarming” given how many people they represent in the court system.
Northrup also described seeing cases in which charges were filed “based solely on an FRT hit,” using the acronym for facial recognition technology, including “a rash of those earlier this year in Baltimore City” that “fortunately” were “resolved.”
Both Baltimore Police and the State’s Attorney’s Office responded to questions about those cases with written statements saying facial recognition technology alone cannot provide probable cause.
Police spokeswoman Lindsey Eldridge said the technology “can assist when coupled with other investigative leads and proven investigative systems and techniques.” She said charges were not filed “based solely on an FRT hit,” but did confirm the agency had received 811 facial recognition match reports last year.
The State’s Attorney’s Office added in a written statement that facial recognition technology is a “valuable investigative tool,” but that it “should not contribute to probable cause for searches or arrests, nor should it be admitted as evidence in criminal prosecutions.”
Several members of the public who testified Wednesday voiced support for the measures as a step toward curbing the use of such technologies.
Martaze Gaines, from Organizing Black and the Campaign for Justice, Safety and Jobs, argued that facial recognition technology and other surveillance tools would “heighten discrimination in policing,” rather than make Baltimore safer.
Gaines added that “every dollar spent on high-tech surveillance is a dollar not spent on community improvements that do much more to increase community safety,” such as school improvements, grocery stores, consistent trash and recycling pickup, and more accessible sidewalks.
Baltimore has made headlines for its use of facial recognition and surveillance tools. Baltimore Police used facial recognition to try to identify people “involved in criminal wrongdoing” during the 2015 unrest following the death of Freddie Gray, a spokesman said in 2016. Around that time, it also faced complaints about using a “stingray” without proper licensure and without disclosing information about it to prosecutors or judges.
The city also flew a surveillance plane, since grounded, for a six-month period in 2020 — until federal judges ruled against the city in a lawsuit and said use of the surveillance power, left unchecked, could violate the Fourth Amendment.
Burnett said Baltimore Police have cooperated in working on his proposals.
The city had a moratorium on facial recognition software (except for Baltimore Police), passed by the City Council in 2021, that expired late last year.
There was some discussion Wednesday about an expected legislative push in the Maryland General Assembly to create statewide regulations. A bill last session would have laid out when police could use the tools, restricted what images could be evaluated as comparisons and required training.
Burnett’s proposals could be an early test of the city’s control over its own police department, which legal officials have described as a gray area. City voters could face a second ballot question on the topic in 2024 — this time, to create a structure for the department within the city charter, to finalize the move from state to local control.
Baltimore Sun reporter Cassidy Jensen contributed to this article.
©2023 Baltimore Sun, Distributed by Tribune Content Agency, LLC.
