It’s a blow to facial recognition made in Clearview AI. In the United States, an association has managed to restrict the playing field of this company specializing in unbridled biometrics. As part of an agreement, this very controversial company is no longer allowed to access certain commercial markets.
This does not sign the final cessation of facial recognition, but it is a very hard blow that has just been dealt against one of the most advanced companies in this field. In the United States, the powerful American Civil Liberties Union (ACLU) succeeded in forcing Clearview AI to stop selling its biometric databases to companies across the country.
The facial recognition is a biometric process, such as the analysis of fingerprints, voice or iris. It is a question here of relying on the features and the shape of the face to recognize an individual.
As part of this agreement, concluded on May 9, 2022 between the two parties and which must still go through a court to be fully valid, Clearview AI no longer has the possibility of dealing with the private sector or with individuals. The company, however, retains the right to work with the authorities – the only exception: Illinois, which benefits from a suspension for the next five years.
A controversial company in facial recognition
Clearview AI is a startup whose activities have come to light in a New York Times investigation. The newspaper showed that the company captured more than three billion photographs from the web in an effort to beef up its algorithms for facial recognition. Objective: then provide a solution to the police, but not only.
This company, which received funding from entrepreneur Peter Thiel, who can also be found in the history of Facebook, Palantir or PayPal, has a database containing more than ten billion faces today, which were taken from the net. It would even be on track to multiply this number by ten, to reach 100 billion by 2023.
Since the NYT investigation, legal actions have begun to emerge across the Atlantic. Even Twitter got involved, as it was noted that Clearview AI dove into the social network’s pages (since it’s in essence mostly public). The subject also took a more political turn, when members of Congress seized on it.
The fact is that all this bad press, risk of lawsuits and legislative tightening has not yet been enough to stop Clearview’s activities – this, while its image has been tarnished by the theft of the list of its clients, its proximity to the conservative milieu and its sometimes somewhat Orwellian projects. But that could change.
” The main provision of the agreement prohibits the company from selling its database of facial prints not only in Illinois, but also throughout the United States. It will also stop offering free trial accounts to police officers without their employer’s knowledge and consent. », summarizes the ACLU on TwitterMay 9.
Towards a federal law to restrict abuse with facial recognition?
If the deal concluded between Clearview AI and the ACLU still leaves a relative room for maneuver for the company specializing in facial recognition, which was undoubtedly inevitable to hope to snatch an agreement, it significantly restricts its commercial prospects in the United States. . Above all, it indirectly highlights the need for more protective legislation in the country.
It was indeed possible to snatch a deal with Clearview AI because the firm was under threat of a more costly and risky lawsuit in Illinois. As the ACLU notesthis state has a law “ innovation that assures residents that their biometric identifiers — including their faces — will not be captured and used without their knowledge and permission. »
Clearview, of course, says get away with it. It is even a huge victory According to the company’s lawyer, quoted by AFP. ” This does not change its business model. “, according to him. She goes ” continue to expand its offer in accordance with the law », and the costs generated by this conflict with the ACLU are always less than what a lawsuit would have cost.
The United States, following in Europe’s footsteps on digital legislation?
This satisfaction could only be a facade. In the longer term, it is not just Illinois that could adopt more demanding legislation regarding the use and protection of biometric information. In a tweet reacting to the action taken by the ACLU, Democratic Senator Ron Wyden welcomed this first step and called for more.
” Good job from the American Civil Liberties Union. Now Congress must finish the job and pass my Fourth Amendment Act to stop government agencies from abusing this dangerous facial recognition system and protect Americans once and for all. », he wrote on May 9 on Twitter.
Another hoped-for benefit of this agreement, pending a hypothetical federal law: it could increase the pressure on all the other companies similar to Clearview AI, by forcing them to further temper their ardor in terms of facial recognition, under penalty of having to in the face of action by the ACLU or another civil rights association.
The ACLU, in any case, considers the pressure of a lawsuit on Clearview in Illinois ” demonstrates that strong privacy laws can offer real protections against abuse “. The company preferred to give in in this state and back down nationally, rather than go to court. ” Other states need to pass similar laws », concludes the ACLU.
In this respect, Europe can serve as a model. In recent years, many texts have emerged on the Old Continent to organize and sometimes restrict the activities and practices of companies in the digital world. We saw it with the GDPR, we see it today with the ePrivacy regulation, the DSA, the DMA, the Data Act proposal or in terms of artificial intelligence.
Clearview, moreover, is in the process of noticing it. In Italy, the personal data protection authority fined him 20 million euros for having collected biometric data on the Italian population. And in France, the Cnil is also on the spot. At the end of 2021, the authority gave Clearview formal notice to stop collecting the photos on the net.