Facial recognition in surveillance keeps growing, but so do doubts
Several deployments of new face biometrics for surveillance reportedly have failed to deliver on crime-fighting promises.
The first is the Bhopal Eye project in Madhya Pradesh, India. According to a new report by United Kingdom-based non-profit No Tech For Tyrant (NT4T), police are using public facial recognition to prevent and solve crimes, “building on the myth that more police equals more safety.”
Instead, the report says biometric surveillance is amplifying historical discrimination perpetuated by law enforcement in the area. Additionally, laws and legal institutions are not adequately protecting citizens against alleged surveillance abuses.
The report also suggests the collaboration between government and technology vendors exacerbates negative surveillance dynamics, particularly insufficient accountability.
At the same time, the authors behind the report said they are optimistic that public resistance is slowly building along with legal challenges.
The NT4T report lists recommendations for specific groups within government and industry. In particular, the non-profit called for researchers to stop recommending intrusive surveillance and repressive policing.
Vendors should take an abolitionist approach to their biometric surveillance development. Finally, civil society should campaign for greater democratic oversight and community accountability when police are authorized to use digital surveillance.
UK’s face surveillance numbers under scrutiny
The Metropolitan Police scanned the faces of roughly 125,000 people in London between February 27, 2020, and July 16, 2022, but made only nine arrests.
The figures come from an Evening Standard investigation that alleged biometrics was used in five deployments, all of which were announced to the public.
A January 28, 2022, operation scanned 12,120 faces, drawing a watchlist of 9,756 people, generating 11 alerts and prompting four arrests.
Three other operations conducted in July 2022 scanned 105,066 faces and prompted the Met Police arresting four people. During these operations, 20,159 people were on the watchlist, and eight alerts were triggered.
Law enforcement told the Evening Standard that all biometric data captured during the trial operations was encrypted in a USB stick. All face information that did not trigger an alert was reportedly deleted immediately.
Biometric data used to import the watch list is deleted as soon as possible (and never later than 24 hours), while other data of suspects is retained for up to 31 days.
Also in the UK, a recent freedom of information request spotted by the Sunderland Echo showed that there were 106 cameras controlled by the Sunderland City Council this summer, an increase of 34 percent compared to 72 in 2019.
The same document also suggests the number of cameras in the UK has increased 15 percent, from 79,022 to 91,081 over the last three years.
NYPD deploys Ring Neighbors
The New York Police Department announced plans last week to deploy Amazon’s Ring Neighbors app, enabling police to access Ring camera video data posted by users and to request specific videos.
According to Audacy, the police are rolling out the technology this week.
To ease privacy concerns connected with the technology, police officials told the publication it will never “have access to customer live streams, and do not have access to customer information, device location or videos unless a customer expressly chooses to share that information.”
Officials told Audacy the technology would greatly help them tackle crime in the city, but a recent analysis by Coda Story seems to suggest the opposite.
Published Thursday, the report suggests several cities across the United States, including liberal bastions like San Francisco, are “turning to authoritarian tech to appear tough on crime,” particularly to secure elections, but also for violent crime.
At the same time, Coda Story reports on the current debate on the repeal of a face biometrics ban for police in New Orleans, saying the argument that biometric surveillance reduces crime is not backed up by empirical evidence.
The investigation also highlights the wrongful arrests of at least three Black men in the U.S. following facial recognition searches. It notes that while public insecurity tends to bring in more support for policing and a willingness to erode civil liberties, this trend may not necessarily be here to stay.
Iran uses biometric surveillance to track protesters
According to the Economic Times, Iranian police are using face biometrics, GPS tracking and social media to track protesters in the country.
The extent to which facial recognition is being used to identify people, and to what extent it is a convenient threat, is unclear. A Scientific American editor looked into the biometric technology, and seems mostly to have found surveillance with other tech tools.
Individuals around the country are also turning to technology to tackle alleged abuses against civilians, mainly women refusing to wear a hijab.
In particular, people are turning to VPNs and an app created by Gershad, a human rights non-profit, that enables users to share and track real-time locations of morality police forces.
Back in Europe, the ban on biometric surveillance is at the center of the AI Act debate, with the new legislation entering its last stages of negotiations early this week.