Facial Recognition AI Violates Personal Freedoms

Again I find another case of AI which tests our ethical limits in terms of the application of this technology and the restriction of personal freedoms and even the violation of human rights.

On this occasion, the issue of the violation of personal privacy will be in the public arena with the launch of a Face Recognition App launched by the startup ClearView AI.

What arouses more suspicion in me is the great effort made by the publicists of the Clearview firm in creating the opinion matrix that their product will help to "identify sexual aggressors." Virtually all the information on the official site is developed around this topic.

Clearview's technology has helped law enforcement track down hundreds of at-large criminals, including pedophiles, terrorists and sex traffickers. It is also used to help exonerate the innocent and identify the victims of crimes including child sex abuse and financial fraud.

_{Quote Source}

Guess where all these images come from?

You got it right! ... Facebook, Google, Venmo, Youtube and other sites are the suppliers of the photographic material.

Current Procedure

The facial recognition technology used by law enforcement agencies is overshadowed by the possibilities offered by Clearview.

Organizations such as the FBI and INTERPOL have been using this technology for a long time but it is limited by certain aspects since contrary to what happens with fingerprints and DNA, which are unalterable throughout life, facial recognition must take into account different factors such as:

Aging
Plastic surgery
Cosmetics
Image quality
Effects of excessive consumption of drugs or tobacco
Subject pose

Therefore, the process must be complemented with a manual phase, making it inefficient.

How does it work?
When a facial image for research is introduced into the system, it is automatically coded using an algorithm that proposes a list of "candidates" of the most likely successes between the system profiles.

Next, qualified and experienced officials manually compare the image under investigation with each of the candidates. They examine the images thoroughly to find unique characteristics that can determine whether it is a "potential candidate", "not a candidate" or "inconclusive".

Data quality
An accurate and effective facial recognition system depends on the quality of facial images. An ideal image would be a passport photograph in accordance with the ICAO standard, since it is a complete frontal image of the subject with homogeneous illumination on the face and a neutral background.

With Clearview AI

We cannot deny the fact that this technology in the right hands would become a valuable tool in the fight against crime.

With only one image captured with a mobile phone or road safety camera, this application can scan all the faces present and compare them with a Big Data of more than 3 billion images, which far exceeds the databases used by the police and the FBI based on passport and driver's license photos, with more than 641 million images of US citizens.

Clearview's image search technology has been independently tested for accuracy and evaluated for legal compliance by nationally recognized authorities. It has achieved the highest standards of performance on every level.

Privacy

There have been many cases of public complaints that we have known about facial recognition:

In March 2016, he won a trial against him where prosecutors alleged that Google Alphabet Inc. collects biometric data from photographs that use facial recognition software and stores them. Prosecutors demanded compensation of more than five million dollars.
Facebook stopped the implementation of facial recognition due to users' concern for their privacy.
Facebook used its users as a great "Mechanical Turk", applying this technology on the photographs that users were uploading and linking identifiable faces with social network profiles so that it was capable, when a user uploaded a photo, of suggest labels with the names of the people who appeared on it; all of this, with the intention of increasing social interaction and the feedback of the facial recognition algorithm itself.

_{Image Source}

Now, ClearView AI has allowed more than 600 law enforcement agencies to use its AI application, although the company declined to provide a list of these agencies. The computer code underlying your application includes programming language to match Google’s augmented reality glasses; allowing users to identify each person they saw. The tool could identify activists in a protest by revealing not only their names, but also where they lived, what they did and who they knew.

Summary

At first glance this technology developed by Clearview IA would be a powerful tool in the fight against crime.
Precisely for this reason all your advertising revolves around this theme. The company says its technology is only available to "law enforcement agencies and companies." This concept results in a very high level of discretion.

How is the KYC performed for granting access to this technology?

On the Clearview site we can see that only 7 questions are asked of the applicants.

Facial recognition will continue to pose major threats to civil rights and freedoms, in the absence of legislation, facial recognition could be used to catalog people who attend religious ceremonies or political demonstrations. The uncontrolled implementation of this technology can exponentiate ethnic discrimination, xenophobia and racism..