This Tuesday, June 9, amid the wave of protests after the death of George Floyd during an arrest in Minneapolis, IBM CEO Arvind Krishna announced in an open letter to the United States Congress that the company will stop marketing software own facial recognition for its rejection of the use of a technology that can contribute to the promotion of "racism and social injustice." Less than 24 hours later, Amazon communicated that for a year it would not let the police use Rekognition, su image recognition platform – although it did not specify anything about its controversial use by the immigration service at borders and customs. On Thursday afternoon it was the president of Microsoft who announced that he will not sell his surveillance technology to the police, while confirming that it had never been sold to him. But these contradictions or advertisements, which seem to have at least as much marketing as they are true engagement, are not new in this area.
Facial recognition is a system that allows a person to be identified by analyzing biometric characteristics of their face such as the distance between the eyes or the size of the nose. There are different technologies that make use of different parameters and with very different precision. Although it was born almost 60 years ago, in the last 10 the use of artificial intelligence and machine learning has exponentially improved identification capabilities in both photos and videos in real time. These technologies are present in the unlocking of many mobiles and in a multitude of apps that put filters or help to tag our photos, but their use by the security forces is the most controversial due to the unprecedented power it gives them to monitor and follow the trace of each citizen.
Earlier in the week, in an interview on The New York TimesTimnit Gebru, who leads Google's artificial intelligence ethics team, asked to temporarily stop the use of these technologies. A surprising statement, especially since it was almost 10 years ago that the then president of the search giant announced that the company had scrapped the investigation on facial recognition as perverse and dangerous. Despite the announcement, Google has continued to invest in its development, as all those who are tearing their clothes this week will surely do. And it is that the technology giants have been in a war for years for a millionaire business with too many open questions about civil rights, social injustice and justification for violence. Its use by the security forces, which some now say they want to give up, is a tiny part of a huge cake: bringing facial identification to companies and consumers; the huge market for facial surveillance of each and every one of us. Let's do some history.
May 31, 2011. Many of Silicon Valley's billionaires gather at AllThingsD’s D9, a gathering organized by Wall street journal at Rancho Palos Verdes in California. On the main stage, Eric Schmidt, CEO of Google. The most powerful man in the most powerful company of digital capitalism opens a debate that lasts years later. Schmidt warns of the challenges of facial recognition technology and the "perverse" implications that its combination with mobile location can have. It warns of the risk that a tool like this could pose in the hands of dictators who use it against the population. For all this, he says, "it is the only technology that Google has developed but later discarded."
Facial recognition is the holy grail that would allow internet giants to jump into the physical world
Nine years later, Eric Schmidt is no longer president of Google, AllthingsD no longer exists – although billionaires continue to congregate on ranches – and Google, like all Internet giants, invests millions of dollars in this "questionable" technology. These investments, those of governments from half the world and those of dozens of startups Backed by large investment funds, they have seen facial recognition capabilities dramatically improve in recent years. The reason seems obvious. Bring together the two most powerful interests in the tech industry. On the one hand the military and security interest – the Chinese Government has deployed Dragon Fly Eye capable of identifying more than 2,000 million people on the street -, on the other hand the economic interest as it is the holy grail that would allow the giants of Internet jump to the physical world. If they are already capable of knowing everything we do in the digital environment, this could be the ultimate weapon to do it in the real environment. We will never be anonymous citizens again.
Months before Schmidt's claims, Facebook had started using facial recognition to suggest the names of the people in the photos and be able to tag them. In 2012, it was forced to stop using this technology in the European Union due to problems with regulation, but in 2018 it implemented it again. For 8 years he used it by default in other areas of the world until in December 2019, pressured by criticism of his privacy management, he added the possibility that the user can remove the permission for the network to use these algorithms with their photos (by default the platform has permission and it is the user who can prohibit it by entering the privacy settings of their profile).
But the Facebook experiments, which was born as a repository on-line of photos of the faces of college students, with facial recognition have gone much further. Between 2015 and 2016 it developed a app that allowed people to be identified with the mobile camera and associated with their profile on the social network. According Business Insider, Zuckerberg's company developed and tested an application for smartphones that allowed to identify people in real life, in order to show their profile on Facebook and some of their information. This application made use of facial recognition based on the millions of photographs of its users stored on its servers (in which if you had profiles on that network between 2010 and 2012, or as of 2018, there is also your face). The company ensures that app was a test within its innovation program and that it only worked to identify employees but it shows that Facebook has the same capabilities to link people in the physical world to their digital data as ClearView, the tool whose use by the forces state security in the US has sparked controversy. These experiments can be very expensive. In January 2020, Zuckerberg's company reached an agreement to avoid going to trial over a $ 35 billion lawsuit by the State of Illinois and a group of users who claimed they never gave consent for Facebook to use their biometric data. .
The death of anonymity
Google is also facing a similar lawsuit in the same state. And, despite what Eric Schmidt claimed, the company has been working with facial recognition for years. Months after Schmidt's remarks, he was introducing Facebook-like capabilities to his defunct social network Google+. The service was called Find my face and scanned photos of users and their friends to identify familiar faces. The Mountain View company's efforts in this area have continued. Its new smart display, the Google Nest Hub Max, includes a controversial feature: It's always looking. Face Match, the name that Google gives to this technology, identifies the people who pass in front of the screen. When it recognizes a registered user, it shows personalized content just for him: photos, messages, quotes. Using similar screens in settings such as streets or shopping centers could achieve an effect very similar to what we saw in the futuristic dystopia of Minority Report. Outdoor advertising, supermarket shelves or offers in personalized physical stores like today are done on the Internet. A millionaire business that nevertheless threatens to kill our anonymity.
Amazon, of course, is also no stranger to this technological battle. Since 2016, Rekognition has included among its cloud services offering, which "democratizes" the use of artificial intelligence, allowing any user to use this technology. The controversy has accompanied the service from the beginning, but in the last year and a half it has received a wave of protests, including from its employees, when it became known that the Trump administration could use Amazon services in different government agencies such as ICE, the US Customs and Immigration Service that Rekognition would use to identify illegal immigrants. ICE has been the protagonist of numerous scandals in recent years and accused of violating the rights of minors and migrants. In a letter, Amazon employees compared the business that their company has with this agency with the sales of technology to Nazi Germany that IBM made in the 1940s.
And Apple? The mobile phone giant that completes the tetrachy of digital empires has also invested in facial recognition. But with a technology very different from the rest, which is not based only on image analysis but also uses infrared. The Cupertino solution is called Face ID and it was released with the iPhone X in 2017 (in fact during the presentation of the phone the technology failed and on stage, with thousands of people following the event live, the presenter was unable to unlock brand new iPhone). Face ID combines the use of the front camera, an infrared camera, a support light and an infrared projector to achieve that the scan is not in two, but in three dimensions. It emits 30,000 invisible infrared points that help reconstruct the final mathematical model of the face, a three-dimensional map that is collated in real time using algorithms based on neural networks and artificial learning capable of processing facial variations. All this allows to recognize faces in situations of absence of light. If Google's Face Match doesn't stop looking, Apple's Face ID is capable of seeing in the dark.
Given these advances, their economic potential and the enormous challenges they pose for privacy, there is no legislation that protects the rights of citizens against these giants and other players that are rapidly entering this market. The European Commission, which had proposed a moratorium that would ban the use of facial recognition in the next five years, has backtracked and softened its position. It will leave it to the Member States to decide how to reinforce that increasingly thin line that protects civil rights. Meanwhile, someone in Silicon Valley may have gotten his face.