Citizens of the European Union would have explicit rights over the use of their facial recognition data under planned regulation by the European Commission.
The aim of the regulation would limit the “indiscriminate use of facial recognition technology” by companies and public authorities, according to an unnamed official who spoke to the Financial Times based in London.
A document seen by FT said the EC wants legislation that would “set a world-standard for AI regulation” with “clear, predictable and uniform rules” and that targets facial recognition with its “specific risks.”
The question then becomes: would such regulation ever fly in the U.S? Several U.S. cities, including San Francisco and Berkeley, California, and Somerville, Mass., have outright bans on use of facial recognition.
The states of New Hampshire and Oregon prohibit facial recognition technology only on body-worn cameras.
Still, the idea of a broad-based ban across the U.S. and all 50 states may seem far off, if ever likely at all. Both the U.S. House and Senate have held oversight hearings, but no concrete measure has materialized.
Democratic president candidate Bernie Sanders, however, wants to ban the use of facial recognition software for policing, according to his recent criminal justice reform plan announced on Aug. 18.
Some U.S. cities have some form of facial recognition technology in place or in the works. In other cities, officials say they only use cameras to watch the movements of crowds to detect breakouts of violence or other emergencies in order to assist first responders. Some cities specifically forbid police from enabling software that can use biometric information from a person’s face, but will still use video cameras to watch for vandalism and other criminal activity near a known crime area. Some video feeds read license plates to match with records of offenders of vehicle laws.
An array of software products for facial recognition are on the market and most use high definition cameras to view a face, then rely on the software to judge the distance between a person’s eyes or the width of a person’s mouth (or a combination of several biometric measures) to compare those findings with a database of known offenders. A big ethical question is what database will be used for comparison, and whether that database is restricted to convicted felons or possibly just a list of arrests without convictions.
Some police and security forces have sidestepped the facial recognition controversy by focusing on use of software that can spot threats in a crowd, such as a person carrying a gun. In one example, police serving the Virginia capitol in Richmond are conducting a test of low-power radar imaging with AI software to detect weapons and threats at the perimeter of the capitol grounds.
In Oakland, California, the city council voted unanimously on July 16 to ban facial recognition software. A report from the council president questioned the lack of standards around its implementation and the potential that the software could be less accurate for women or dark-skinned people.
“Decisions about whether we want to hand the government the power to identify who attends protests, political rallies, church or AA meetings should not be made in the secret backroom of a police station, lobbied by corporate executives that market this technology,” an American Civil Liberties Union official said in an email to Fox News.
The ACLU recently tested Rekognition, facial recognition software from Amazon, finding that about one in five California legislators were erroneously matched to a person who had been arrested. The ACLU used the software to screen the legislators’ pictures against a database of 25,000 publicly-available booking photos.
Some law enforcement officials defend facial recognition technology as an important tool for law enforcement in policing crowds or to find lost children.
Bill Bratton, former police commissioner for New York City, commented recently that combining private and public security video feeds can be especially valuable in a quick response to a crime or fire. “In times of terror attacks and mass shootings, we can get our arms around these problems in a faster, more secure way,” he said on a recent webinar about smart city technologies. But even so, he added that when it comes to using facial recognition on top of video feeds in cities, privacy issues “need to be resolved.” He didn't offer a specific remedy.
The use of facial recognition is likely to increase dramatically as more and more cameras are installed in cities, said analyst Jack Gold, at J. Gold Associates.
“We need to have some form of facial recognition standards for privacy and security,” Gold said. “It’s likely that other parts of the world, including Europe, will lead in the regulatory space. The risk of having my facial data online somewhere for people to access could allow hackers to use that to access my bank accounts, travel, healthcare records and other very personal information or financial assets. It is very troubling. “
“Or, what happens if someone decides to replace my facial data with someone else’s? There’s no current standard that I’m aware of, but we really need to get ahead of this one,” he said.
Gold said he suspects the U.S. “will drag its feet for a long time until something disastrous happens and before anything gets done. I’m not very optimistic this will get the oversight that it truly needs, either from tech companies themselves or from regulators.”