Face recognition: banned by IBM and Amazon. Here is why

Face recognition: banned by IBM and Amazon. Here is why

An accusing finger has been pointed at face recognition technologies which are being blamed of being racist and of having been used improperly. So, after the crowds which have invaded the US streets, also the Big  Techs are getting into the way. Here is what has been happening in the West of the world, and the Chinese exception.

Face recognition technologies are being banned. After  George Floyd’s murder and the consequent riots against the police which have exploded around the world, even the Big Techs are supporting the organizations defending civil rights and privacy, backing up angry citizens.  What is being asked is a law regulating the use of facial recognition technologies, stating when and why they can be used.

Facial Recognition on trial

The main – and also the most serious charge – is that facial recognition technologies can easily be “used and overused” and that they may be employed for ethnic profiling thus being essentially racist. Yes, you have read it correctly, I said racist, because the AI algorithm behind such technologies apparently cannot identify faces with a dark complexion, which means that only white people may be excluded, while very different ethnic groups such as Asians, black people and Latinos would be thrown into the same pot.

And again: they reinforce prejudices, act in breach of the privacy law, undermine civil rights and may generate mistakes concerning the identification of criminals, since, once the software has identified a suspect with a level of certainty over 90%, investigators may be misled to believe something untrue. We are talking about a controversial technology, then; still, it is being used almost everywhere. Also in Italy (Why Como has been one of the first cities in Italy to use face recognition software)

IBM, Amazon and Google: thumbs down concerning Face Recognition software

IBM was the first company to take a stand against face recognition software. The company CEO, Arvind Krishna, sent a letter to the United States congress stating that the company strongly opposed to the use of facial recognition technologies and would not accept its application as an instrument for mass surveillance, ethnic profiling, breach of human rights and freedom (Facial recognition and ethnic prejudice: Amazon and Ibm against the US police).

All the other tech giants have followed since then, even those which actually sell such technology such as Amazon, whose Rekognition, a highly controversial facial recognition software, has been used by governmental agencies. Considering the situation and the most recent events, Bezos decided to adapt and stopped the use of such technology by the police for a year, waiting for the Congress to issue a resolution thereof. Only exception, the application of such technology to tackle human trafficking and to find missing children.

Timnit Gebru, leader of Google ethical commission for Artificial Intelligence has also taken a stand against facial recognition which, in his opinion, seems to be less accurate than the process carried out by human beings, may be detrimental to some communities and has already been used for profiling protesters. In short, this technology is likely to have been used too early, while still being imprecise and not transparent enough (A case for banning facial recognition).

Not to mention the psychological aspect of such issue, which is not strictly connected to this technology in particular. Humans, in general, tend to rely on technology too much and, consequently, we stop checking, thinking, being critical.

The ClearView AI case

Concerns (legitimate) regarding facial recognition apparently do not apply to  ClearView AI. We are talking about an app, developed by an American start-up, designed to search the net and collect users’ public images without notifying them. At the moment, the company claims to have put together a database of about 3 billion images which,  if compared, may allow individuals to be identified with a level of accuracy of 99,6%. The app should have been reserved for police corps and governmental agencies, but over the past few months, rumours have spread about it having been sold to single individuals willing to pay big amounts. Actually, despite the fact that the producer is not famous for its transparency, it seems to have several countries among its loyal customers and is allegedly trying to conquer Europe, and also Italy, where it is anyway facing an escalation of the shields and – luckily for us – the GDPR (A controversial startup producing facial recognition software wants to make business with Italy).

For the company patron, Hoan Ton-That, business is business. And, in any case, facial recognition is their core business. So, arguing that their software helps decreasing crimes, ClearView not only did not align with the Big Techs but has openly stated they will continue selling their technology to police corps around the world (Clearview AI Won’t Stop Providing Its Tool To Police Departments).

China loves biometrics

It has been known for quite a while: unlike Europe, in China the concept of privacy does not exist. It is no surprise that facial recognition technologies are widely used across the Asian superpower. Facial recognition is part of Beijing safety plan and every 1.000 inhabitants there are over 100 cameras exploiting this technology for controlling people. Currently, facial recognition is also being tested in the underground system and even when it comes to purchasing new Sim cards (Biometric data: 50 countries ranked by how they’re collecting it and what they’re doing with it)

The ban may be useful

The true, sole, big purpose of technology ought to be improving our lives. Therefore, should its role become not only controversial but also potentially detrimental, we  need to stop and take the time to think it over and work out some specific and clear guidelines. I hope politicians will act accordingly within one year, or maybe less.

What do you think about the facial recognition technology? In which cases do you think it may be useful and rightful to use it? Tweet @AgostinelliAldo

If you liked this article, you may also like: “Marketers and AI: investments are increasing. Here is why