It ended up in the crosshairs of the British, French, Canadian and Australian authorities. Contested by civil liberties and privacy associations. Subject of journalistic inquiries and appeals as well as of lawsuits, especially in the United States. The open fronts around Clearview AI are many but the controversial company of Facial recognition goes on as if nothing had happened, also betraying the promises of a couple of years ago on the sale of its services to non-governmental bodies (it has not yet decided whether to sell them to them or not), and aims to collect within the year 100 billion images.
The latest revelations from the Washington Post
According to the revelations of Washington PostIn fact, in a communication to its investors, the group – of which little or nothing was known until 2019 – would be able to achieve this monumental goal within the year, if however it receives the additional funds it asks for: a 100 billion file face image database. Practically how much would be enough to make sure that “almost everyone in the world is identifiable“, as a presentation of a couple of months ago obtained by the US newspaper reported verbatim. But, in fact, another 50 million dollars are needed.
What is Clearview AI
Clearview AI is used by thousands of agencies and law enforcement agencies around the world in operations, investigations, investigations and other activities and allows you to compare images held by an operator with those in the company’s powered database. Even the Italian State Police – which basically uses Sari, Automatic Face Recognition System – would have used the platform on many occasions, especially in the early stages after the service’s landing in Europe in 2019, when the company offered 30 days of free tests. . This proved an investigation some time ago by BuzzFeed News, following which – last September – Deputy PD Filippo Sensi presented a question on the subject to Interior Minister Luciana Lamorgese.
According to the group their own database would be the largest of its kind, that is, considering the faces of human beings, and would exceed by eleven times any existing archive of biometric data in the hands of other entities, governmental or otherwise. March to the rhythm of 1.5 billion images added every month and the disturbing aspect is that they are photos that we mostly providedirectly, by uploading them to social networks, blogs, sites and other platforms.
How Clearview AI builds its database
Yes, because despite the opposition of giants like Facebook, Google, YouTube or Twitter, Clearview AI populates a large part of its database from trawling carried out on social networks. The writer last year made a request for access to his images (you can do it by following the instructions indicated here, providing a photo through which Clearview AI returns a report of what he has in his hand), finding several dozen of them. And asking for it removal from the database. You also need an identity document.
Clearview AI does indeed data scraping, that is, it automatically collects images from thousands of sources, including social media, simulating the browsing activity of a normal user thanks to automated systems and on a huge scale, so as to avoid any security measures implemented by content managers. At the beginning of 2020 the database counted three billion photos, today they are already 10 billion and the pace is precisely that seen above.
What does the First Amendment of the US Constitution have to do with it
When someone objects that the authorization of the parties concerned or of the companies is lackingClearview AI pulls the First Amendment to the US Constitution. This was the case, for example, when the group was sued by the American Civil Liberties Union for violating the Illinois Biometric Information Privacy Act, a local state law. According to the company, founded and led by Hoan Ton-That (born in Australia to a Vietnamese family and moved to San Francisco), the collection of publicly available information, the analysis of this information and the sharing of the conclusions of this computer processing would make the platform comparable to a search engine. An interpretation certainly beyond the limits of that perimeter, which the US Supreme Court has also greatly expanded in recent years. And indeed, just last Monday, a federal judge did rejected that line of defense based on the exploitation of the First Amendment and allowed the process to continue: “A major victory for our privacy over Clearview’s profits,” wrote the Electronic Frontier Foundation.
Beyond facial recognition
In short, Clearview AI is seeking further funding – for this purpose the document told by the newspaper was needed – to push on lobbying towards politicians around the world, so as to favor sufficiently large privacy laws, develop new products and expand the international sales team. The prospect is to march without paying any attention to criticisms, objections, investigations and doubts e propose tools far beyond facial recognition, to be able to identify people all over the world through various other parameters obtainable from the algorithms that analyze the images in its possession: from geolocation to the way of walking up to being able to obtain the fingerprints of the subjects immortalized remotely. All this not only from desktop or mobile workstations but directly through augmented reality glasses on which it is carrying out tests and researches and that in perspective the police officers will be able to wear to carry out a query of the database practically in real time during the operations.
With respect to the revelation of the Post, the company responded with brief press notes, returning to insist on the comparison with search engines. This, for example, the one sent to the newspaper Ars Technica: “Clearview AI’s database of publicly available images is legally collected, just like any other search engine, including Google. It is used by law enforcement for post-fact investigations to help identify the perpetrators of those crimes. “In addition, to the Post, the founder added that” every photo in the database is a potential clue that could save a life, provide justice to an innocent victim, prevent illicit identification or exonerate an innocent person “, also recalling how Clearview AI was used to investigate the assaults on the US Capitol.