Police in Miami, India, Macau expand facial recognition use amid further warnings
The Miami-Dade County Police Department (MDPD), Florida’s largest police force, is seeking to join the many departments in the state with access to the facial recognition database operated by the Pinellas County Sheriff’s Office (PCSO), which is one of the largest in America, Miami New Times reports.
Documents filed with the Public Safety and Rehabilitation Committee of the county show MDPD is attempting to reach a memorandum of agreement with PCSO to use its “FACES” (Face Analysis Comparison and Examination System).
“Technology drives much of today’s innovation and advancements in law enforcement,” Miami-Dade Mayor Carlos Gimenez wrote in a memo to the public-safety board. “It also provides the foundation for day-to-day operations such as criminal background histories for critical information sharing. The Pinellas County Sheriff’s Office maintains the largest collaborative, open-model face recognition system in the United States and exclusively hosts Florida’s Facial Recognition Network (FR-Net).”
The MDPD plans to use the system to perform automated facial recognition searches, and to help forensic artists create composite sketches. Capt. Gustavo Duarte, who oversees the department’s use of body-cameras, will be in charge of its use of facial biometrics.
Miami New Times reports that 243 police departments across the state were using FACES as of 2016, that it is used for about 8,000 searches per month, and that the system’s accuracy has been subjected to minimal auditing.
Meanwhile in the U.S., non-government organization Open the Government has joined forces with MuckRock to send dozens of FOIA requests to police departments to gather information about their use of facial recognition under its new “Police Surveillance: Facial Recognition Use in Your Backyard” project.
The project has so far found that Hawaii and Ohio are among states running their own state-wide facial recognition systems, while Clackamas County, Oregon, has been experimenting with Amazon Rekognition in coordination with other agencies, despite a claim to the contrary by its public information officer earlier this year.
New implementations in India
The National Crime Records Bureau (NCRB) of India says its incoming facial recognition system does not violate the principle of consent, and will be operated according to strict safeguards, according to the Financial Times.
Privacy advocates have worried that the system lacks an appropriate legal and policy framework, but in response to a challenge by the Internet Freedom Foundation, the NCRB has stated that the system will only be used in criminal investigations, and available only to police through a secure environment.
As the tender announcement for the system is awaited, the Railway Protection Force (RPF) of Indian Railways is planning to link a facial recognition system at rail stations around the country to the Crime and Criminal Tracking Network & System (CCTNS) and other databases to identify criminals and suspects, Indian publication Swarajya reports.
The RPF drew up plans to install facial recognition at 200 rail stations following the 26/11 terrorist attacks in Mumbai. Activists have expressed concern at the system’s potential impact on privacy.
“We, as a proof of concept, began working with facial recognition technology at the Bangalore station and plan to implement it slowly across our entire network,” says RPF Director General Arun Kumar.
The Indian state of Uttar Pradesh has deployed a video analytics platform from startup Staqu which features facial recognition to analyze feeds from 700 CCTV cameras across 70 prisons, according to Inc42.
Staqu’s Jarvis platform also includes crowd analysis and unauthorized access alerts. The company has previously developed an app called Trinetra for law enforcement in several states to identify suspects from a central database of criminal convicts with facial recognition.
UK commissioners concerned
The Jersey Office of the Information Commissioner has expressed serious concerns with potential risks to privacy from the use of facial recognition, echoing a call by the UK Information Commissioner for a binding statutory code of practice to regulate its use by law enforcement.
An announcement from the Jersey Commissioner’s Office notes that the Data Protection (Jersey) Law 2018 (DPJL) designates personal information, including biometric data, as a special category of data.
“FRT processing involves a high risk to the rights and freedoms of natural persons,” states Jersey information Commissioner Dr. Jay Fedorak. “Studies have demonstrated that it is prone to error, including false positives. And these errors are greater for individuals of different ethnic backgrounds.”
Due to these risks, Fedorak says data protection by design and default must be incorporated into facial recognition systems, and data protection impact assessments are necessary in situations with a high risk to rights and freedoms. Because of this, Fedorak urges Data Controllers to consult with his office before processing high volumes of special category data.
Like all video surveillance, Fedorak says, facial recognition should be used as a last resort.
The North Wales Police and Crime Commissioner Arfon Jones has compared the use of facial recognition at soccer matches in the country to “Big Brother,” North Wales Live reports, after it was used at a recent match where some fans wore Halloween masks and called for stronger controls over the technology.
“It’s not proportionate to use that technology to take pictures of supporters at football matches,” Jones explains. “To put it into context, there were around 20,000 people at that match and they were worried about the behaviour of around half a dozen who have been banned from football matches.”
“We already know who the troublemakers are and they’ve been banned. We have specialist staff, and all they do is deal with football matches, and they know exactly who to look out for,” Jones continues. “I think it’s more proportionate to deal with these people individually instead of using the Big Brother approach with 99 percent of the population; including children who have had pictures taken of them.”
South Wales Police and Crime Commissioner Alun Michael disagrees, and calls it “incomprehensible” that Jones would not support measures to keep spectators safe.
He further said that the technology has not resulted in any false arrests, but did catch a banned individual trying to enter a recent match.
“This particular deployment was very specific and follows the principle that prevention of violence is better than having to respond after violence has happened,” Michael explains. “That is in the best interests of football fans and the reputation of the sport itself.”
Jones responded that Michael had ignored his concerns, as well as those of UK Information Commissioner Elizabeth Denham, and had missed the point about the thousands of false positives at a soccer game in 2017.
Macau police announce plans
The Special Unitary Police (SPU) of Macau have announced that they will begin testing facial recognition on some of the city’s CCTV surveillance cameras in 2020, according to the Macau Daily Times.
The protectorate’s government previously said that facial recognition would be implemented on at least 100 of its ‘Eye in the Sky’ cameras in 2019, and that the CCTV system would include at least 1,620 cameras by 2020.
Now, police say they will test the technology on 50 cameras in the first quarter of next year, and an additional 50 in the second half of the year. Images will be automatically deleted at the conclusion of the allowable 60-day period. In the event they contain evidence, images will be deleted within 30 days of the completion of the relevant judicial processes. The SPU also says it will consult with the Office for Personal Data Protection prior to the test.
The government now plans to have 2,600 CCTV cameras deployed by 2023, and 4,200 by 2028.
Vuzix and Sword launch facial recognition glasses
Smart glasses and augmented reality company Vuzix has received a $7.1 million order for 10,000 units of its Vuzix Blade Smart Glasses, and reached a reseller agreement with enterprise security company Sword.
The Sword Enterprise solution includes biometric facial recognition and mobile-based threat detection capabilities, and will now include Vuzix Blade Smart Glasses for use by sports arenas, public safety providers, security firms, banks, public schools, airports, federal agencies and telecommunications providers, according to the announcement.
“Vuzix Blade Smart Glasses are the ideal wearable solution to pair with our SWORD Enterprise solution as the see-through smart glasses allow our customers to stay engaged with their environment, while also simultaneously receiving critical alerts related to safety threats from near or across the building from the SWORD detection system,” said Barry Oberholzer, founder and CEO of Sword. “We look forward to a long-term relationship with Vuzix as a smart glasses supplier and technology partner.”