Cambridge moves forward with biometric facial recognition ban as police refine uses
The city council of Cambridge, Massachusetts has approved a bill prohibiting the use of biometric facial recognition by local government agencies for review by the Public Safety Committee, moving it one step closer to a ban on the technology, Gizmodo reports.
If Cambridge were to ban local government facial recognition use, it would likely be the fourth city in the U.S. to do so, after Sommerville, Massachusetts, San Francisco, and Oakland. Following the Public Safety Committee hearing, it would move to an ordinance committee, and then back to city council for final approval. Mayor Marc McGovern, who sponsored the amendment along with two councilors, says that the process will be completed by the end of the year.
“The use of face recognition technology can have a chilling effect on the exercise of constitutionally protected free speech, with the technology being used in China to target ethnic minorities, and in the United States, it was used by police agencies in Baltimore, Maryland, to target activists in the aftermath of Freddie Gray’s death,” the amendment says.
NYPD juvenile facial recognition database
The New York Police Department has loaded arrest photos of thousands of children and teenagers into its facial recognition database, despite the higher risk of false matches with young faces, according to internal records reported by the New York Times.
Children as young as 11 are included, although most are between 13 and 16, the Times reports. The juvenile database includes 5,500 individuals, more than 4,000 of whom are not longer 16 or younger.
Several city councillors and civil liberties groups said they were unaware of the practice when contacted by the Times. A spokesperson for the mayor said he is aware of the department’s facial recognition policy for minors, and that City Hall monitors the departments adherence to “strict guidelines.”
“I don’t think this is any secret decision that’s made behind closed doors,” NYPD Chief of Detectives Dermot F. Shea told the Times, saying the inclusion of the images fit with its regular process. The department’s legal bureau has approved the use of facial recognition on juveniles, but Shea notes that arrests cannot be made solely on a facial recognition match.
University of North Carolina-Wilmington computer science professor and Face Aging Group Co-founder Karl Ricanek Jr. says that performance problems such as racial bias will be exacerbated when trying to match images taken while young people’s faces are still significantly changing.
Idemia and DataWorks Plus, which provide facial recognition technology to the NYPD, did not respond to the Times requests for comment.
State law requires arrest photos to be deleted if the child has only committed a misdemeanor, is cleared, or reaches 21 without a criminal record.
Detroit police change policy
Detroit police have removed a clause from their facial recognition policy that would have allowed for use of the technology with live-streaming feeds in the case of a credible threat of terrorist attack, The Detroit News reports.
Police Chief James Craig notes that the federal government would take over investigations of a credible terror threat anyway.
“Another area we tightened up: the police commission wanted to know specifically what would be the response if someone were to violate the policy,” Craig says. “The new policy calls for dismissal and possible criminal charges if anyone abuses it.”
Police Commissioner Willie Burton was arrested during a police board meeting in July, after expressing concern about the technology, and asking new board chairwoman Lisa Carter if she plans to run the board differently than during her previous stint in 2017-2018. Burton says the police should have gone through a process including consultation with the board before implementing the technology. Craig says it was mentioned in a City Council meeting, but admitted he may have gone about the process differently if he had known what the reaction would be.
Craig has opened up the system to tours by reporters, citizens, and police commissioners.
“I was skeptical, but now I support it,” Police Commissioner Willie Bell said. “(Police commissioners) were impressed. Once you see how this works, it takes all the myth out of it.”
Burton, however, calls the technology “techno-racism,” and questions why Detroit, which he says is America’s blackest city, is “so dead-set” on using it.
Craig says that the process is very rigorous, requiring agreement by two analysts and a supervisor, and that only 30 percent of 500 photos were forwarded to detectives.
Biometric trial boosts Mumbai police
The launch of an automated multimodal biometric identification system (AMBIS) in Mumbai has helped police solve 85 criminal cases and identify 599 individuals accused of crimes during the first four days of a trial, the Hindustan Times reports.
The system, which according to the Times was developed in France, enables biometric fingerprint, iris, and face searchers. The AMBIS system cost a reported 500 million rupees (US$7.2 million), and has the capacity to store 2 million records and return search results in less than half a second. AMBIS records can also be shared with the National Crime Records Bureau, which is in the process of setting up its own Automated Facial Recognition System (AFRS).
“Data of 6.5 lakh (650,000) criminals across the state is already stored in the system till date. And the most interesting part of the system is that even if a criminal changes his looks, the system will catch him/her with its unique and accurate facial recognition system,” said Balsing Rajput, superintendent of police (cyber).
Legal technology company Ipro considers the potential role of facial recognition in eDiscovery in a post to the company website.
Any new technology that creates data must be discoverable for criminal and civil cases, but facial recognition could also be used to sift through other kinds of discoverable material. Veritone has launched an AI eDiscovery tool to allow unstructured data to be searched by keyword, face, or object.