US cities abandon ban on face recognition as crime increases, but critics point to racial bias


In the United States, state and administrative authorities are turning 180 degrees over the ban on face recognition. Several cities and states, including Virginia, New Orleans and California, are in the process of lifting a ban on the use of face recognition by law enforcement. The country will face an increase in crime, and this technology is seen as a way to combat this state of affairs. Critics point to racial bias in face recognition and other technological biases.

Face recognition is returning to the United States as bans aimed at thwarting technology and curbing racial bias in policing threaten to increase crime and increase lobbying by developers. Efforts to impose bans have met with resistance from jurisdictions of all sizes, from New York and Colorado to West Lafayette, Indiana. Even Vermont, the last state to ban police from 100% face recognition, amended its law in 2021 to allow the investigation of children’s sexual crimes.

In July, Virginia will lift the ban on local police from using face recognition only a year after it is approved. Analysts believe that California and the city of New Orleans may press the “cancel” button this month. The reason for such a change on the part of the authorities would be an increase in crime. Reports of killings in New Orleans have risen 67 percent in the past two years than the previous couple, and police say they need all the tools they can get. Other cities show higher figures for the same period.

It will be recalled that Virginia approved the ban due to a process that limited the contribution from face recognition developers. “This year, corporate lobbyists have succeeded in promoting legislation that better balances individual freedoms with the needs of law enforcement,” said Sen. Scott Sarowell. Starting July 1, Virginia police will be allowed to use face recognition software, which tests have shown can achieve 98 percent accuracy with minimal changes based on demographics.

“Technology is needed to solve these crimes and bring individuals to justice,” said Police Commissioner Sean Ferguson. Ferguson asked the city council to lift the ban, which came into force last year. From 2019 to 2021, about two dozen states or local governments in the United States have passed laws restricting the use of facial recognition. Studies have shown that this technology is less effective in identifying black people, and protests against the Black Lives Matter police have given some impetus to the arguments.

But current research by the National Institute of Standards and Technology of the Federal Government (NIST) has reportedly shown significant progress in accuracy. And tests by the US Department of Homeland Security, published last month, revealed slight differences in accuracy in skin color and gender. “There is a growing interest in technology approaches that address technology concerns,” said Jake Parker, senior director of government relations with the Security Industry Association lobby group.

According to Parker, the most important thing is to ensure that it is used in a limited, accurate and non-discriminatory way that benefits communities. The change in sentiment could allow its members, including Clearview AI, Idemia and Motorola Solutions, to get the most out of the $ 124 billion that states and local governments spend annually on policing. The part about technology is not followed carefully. It is becoming increasingly important for Clearview to get new contracts with the police.

This week, the company closed a privacy lawsuit over images collected from social media, agreeing not to sell its flagship system to the US private sector. Clearview, which helps police find matches on social media data, said it welcomed “any regulation that helps society get the most out of face recognition technology by limiting potential downsides.” Idemia and Motorola, which match government databases, declined to comment.

Although recent research has eased lawmakers’ cautions, the debate continues. The Office of General Services, which oversees federal contractors, said in a report last month that basic facial recognition tools could not be disproportionately equal to African Americans in its tests. The agency, however, did not provide more details about the tests. Face recognition will be tested by the President’s new National Artificial Intelligence Advisory Committee.

In Virginia, critics of the standard said it was well-intentioned but false, and that warrants were needed to use face recognition. Combating police discrimination by double-checking the algorithm is like trying to solve police brutality by testing whether a weapon is racist: strictly speaking, it is better than an alternative, but the real problem is who keeps it, ”said Os Keys. , Ada Lovelace Fellow at the University of Washington.

Following the defeat in New Orleans, Virginia, civil liberties groups are growing. Last week, ten national organizations called on city council members to strengthen rather than lift the ban, citing the risk of illegal arrests based on misidentification. The local group Eye on Surveillance said New Orleans “can’t afford to go back.” In the United States, as in France, the use or non-use of facial recognition is one of the most discussed topics in recent times.

In a report presented to the legal commission on Wednesday, French senators suggested experimenting with face recognition for three years. The report’s authors, Senators Marc-Philippe Dobress, Arno de Belene and Germain Duren, said they were against the surveillance company, but suggested “experimenting with face recognition for 3 years in limited, controlled and controlled cases”. The test results should provide a legal basis for the use of this technology. The report contains 30 recommendations around which this experiment should revolve.

However, Recommendation № 3 proposes to prohibit remote real-time biometric surveillance during demonstrations on public roads and around places of worship, except in specific cases. They justify “very limited exceptions” to terrorist risk, the investigation of serious offenses that threaten or harm the physical integrity of people, and other real risks. At a news conference on Wednesday, three speakers insisted that they were against the surveillance society and that their work had helped draw red lines on the use of biometric recognition.

And you?

What is your opinion on this?
What do you think about the use of face recognition in American cities?

See also

France: Senate report recommends experimenting with face recognition for three years and draws red lines on the use of this technology

France: Government intends to launch National Electronic Identity Card (CNIe) from 2021, a program based on face recognition through the Alicem application

France is calling on Clearview AI, an American facial recognition company, to delete its data, and it has two months to enforce the bans.

Clearview AI, a controversial facial recognition expert, has agreed to limit sales in the United States after a high-profile lawsuit.