Enlarge image
Hoan Ton-That, CEO of Clearview AI, demonstrates his software
Photo: Seth Little/AP
Controversial facial recognition startup Clearview AI has agreed to stop selling access to its biometric database to companies in the United States and to government agencies and law enforcement agencies in the state of Illinois for at least five years.
The agreement with civil rights organization ACLU, announced Monday, says Clearview AI can no longer make its facial database available to most private entities in the US, either for free or for money.
However, the company can continue to work with the authorities – provided local laws do not conflict with this.
Government agencies are no longer allowed to use facial recognition in 17 US municipalities, including San Francisco, Oakland and Boston.
However, it must provide Illinois residents with an opt-out offer and promote it with $50,000 worth of online ads.
And it is no longer allowed to provide police officers with free trial accounts without their superiors being informed.
The agreement still has to be confirmed in court.
The ACLU and other organizations sued Clearview AI two years ago for allegedly violating the Illinois Biometric Information Privacy Act (Bipa).
ACLU representative Nathan Freed Wessler said the company "can no longer fully treat people's unique biometric identifiers as a source of profit."
Lawyer: Changes "nothing to the current business model"
Wessler hopes other companies will follow suit.
"This agreement shows that strong data protection laws can provide real protection against abuse," he said.
In the United States, there is no federal law protecting online privacy.
Clearview's attorney, Lee Wolosky, said the agreement "doesn't change anything about the company's current business model."
The start-up saved “a lot of money” through the out-of-court settlement and could “further expand its business in accordance with the law”.
In February, however, an investor prospectus was made public, which revealed that Clearview's technology could also be used, for example, to monitor employees in the so-called gig economy - i.e. people who are temporarily employed for smaller jobs and are often already worse off are as regular employees.
According to Clearview, it has a database that now contains more than 20 billion images of faces that the company has collected on social networks, corporate or other websites without obtaining the consent of those affected or the respective site operators.
These photos are incorporated into the company's identification software.
The next goal, according to Clearview, is 100 billion photos.
(Read more about whether and how to get out of the database as an individual Internet user here. )
Civil rights activists had warned of the dangers of commercial facial recognition providers.
For example, violent partners or sex offenders could use it to track down their victims, even if they are in hiding, warned Linda Xóchitl Tortolero, chair of the Chicago-based organization Acción.
She therefore described the agreement as a "great victory" for people at risk.
Several lawsuits are also pending against the company in European countries.
The Italian data protection authority fined Clearview €20 million in March for enabling "biometric surveillance of persons present on Italian territory".
Most recently, Clearview made its technology available to the Ukrainian army.
They want to use it to identify killed Russian soldiers in order to be able to inform their families, but also to identify living suspects at checkpoints.
Civil rights activists warn that if the software fails, it can have deadly consequences.
pbe/AFP