Facial recognition is an increasingly widespread technology, based on artificial intelligence, which allows a person to be identified in a photo or video by comparing his or her face with those stored in a database. Find out everything you need to know about this technology, how it works and the dangers it poses to privacy…
In 1949, In his 1984 book, George Orwell imagined a counter-utopic society… in which citizens would be under constant scrutiny by the eye of Big Brother. While this work of fiction has fuelled the nightmares of many readers around the world, it did not prevent the world from converging on a similar model nearly 36 years after the date predicted by the British writer .
At the dawn of the 2020s, Facial recognition technology is being adopted for surveillance purposes. in most developed countries. In order to better understand the major turning point that this innovation represents for humanity, find out everything you need to know about its usefulness, its functioning and its dangers…
Facial recognition: what is it and how does it work?
Facial recognition is a technology combining biometric techniques, artificial intelligence, 3D mapping and Deep Learning to compare and analyze a person’s face to identify them. It owes its recent growth to advances in the fields of Big Data, neural networks and GPUs.
Currently considered one of the three most powerful biometric technologies for identifying an individual, facial recognition is also the fastest growing biometric technology. Its market could reach a value of $7.7 billion by 2022..
First, an individual’s face is located on a photo or video. The characteristics of his face are then converted into dataand this data can then be compared with face data entered in a centralized database.
Generally speaking, current facial recognition software analyzes about 80 facial characteristics, also called nodal points.. These characteristics include the distance between the eyes, the length of the nose, the shape of the cheeks, the depth of the eye sockets, and the width of the jaw.
These traits differ on each individual, which is why facial recognition allows us to accurately recognize a person. The points thus collected are measured by creating a numeric code called faceprint.The face can be represented in a database. The most recent new technologies are based on the skin texture, specific to each individual, for even more precise results.
To be able to instantly detect a face, facial recognition systems are based on artificial intelligence. Thanks to Deep Learning, the algorithms are trained to recognize human faces… from numerous photos and videos. Many companies train their neural networks on the billions of photos of faces stored on the Internet by Flickr, Instagram, Facebook or Google.
What are the applications of facial recognition?
This technology is increasingly being used for surveillance and security purposes, particularly by governments and authorities who are tending to incorporate it into video surveillance systems. This is the case in many countries around the world. The China is undoubtedly the main example as the government is seeking to generalize its use.
However, facial recognition is also increasingly used in France, the United Kingdom and the United States.. The city of Nice, for example, is experimenting with facial recognition for surveillance purposes as part of its Carnival, while the US government uses it at airports to identify individuals whose visas have expired.
It is also increasingly exploited by companies in various industries such as health, marketing or tourism. It is also found on many services and products for the general public.
For example, for example.., from the 2017 iPhone XApple’s smartphones feature Face ID technology, which allows users to unlock them by presenting their faces to the front camera. A 3D scanner compares more than 30,000 features to accurately verify the user’s identity. Face ID also allows users to validate purchases with Apple Pay.
On his side, Facebook develops DeepFace technology. It automatically identifies people’s faces in photos posted on the social network with 97% accuracy. Each time a Facebook user is “tagged” on a photo, the characteristics of his or her face are mapped by the system. When enough data is collected, the software is able to identify the person automatically in all new photos.
In June 2015, Google launched its FaceNet facial recognition system. Thanks to an innovative algorithm and an artificial neural network, this system achieves a record accuracy of 99.63%.. It is embedded in Google Photos to automatically tag photos when a face is recognized.
Let’s also mention the FaceApp application, which met with great success in the summer of 2019 to the point of going viral. This app has much contributed to the democratization of facial recognitionby allowing the user to visualize what he will look like when he is old…
For their part, Amazon, MasterCard and Alibaba now offer their users the possibility of confirm payment for their online purchases via facial recognition. This new authentication method is called “selfie pay”.
In a large number of airports, smart billboards are capable of identifying the gender, ethnicity and age of passers-by in order to offer them targeted advertising. This innovation is also based on facial recognition.
Thanks to the giants of technology, Facial recognition is becoming more and more accessible to application developers.. For example, the Amazon Rekognition image analysis service makes it easy to add this technology to an application. This is also the case of the Google Cloud Vision API. Thus, developers can give free rein to their creativity…
The History of Facial Recognition
It’s in the 1990s as facial recognition began to gain popularity…when the U.S. DoD was looking for technology to detect illegal aliens. Leading scientists and experts were then approached to develop this innovation, and their research was funded.
In early 2001, facial recognition made a lasting impression. when first used in a public space at the Super Bowl XXXV in Tampa.. The authorities had then used it to detect possible criminals and terrorists among the spectators. Systems were then deployed to other high-risk areas of the United States to monitor criminal activity.
With the rapid development of Big Data and artificial intelligence, Facial recognition has developed a lot during the years 2010. As a result, its adoption is spreading in a wide variety of sectors for a wide variety of uses.
How is facial recognition used around the world?
It is now estimated that facial recognition systems deployed at U.S. airports and through other processes have captured and stored the facial data of more than half of the citizens of the United States. This information, stored in a centralized database, can be freely explored by the authorities. In more than half of the major cities, police officers are equipped with body cameras for real-time facial recognition.
In Europe, apart from shops and private companies, facial recognition is used far less than in other developed countries. For good reason, the GDMP in force since May 2018 protects citizens’ biometric data. It is not not the case in the United States, where no law currently governs the collection and use of such data; and.
The The worst use of facial recognition is undoubtedly the one China makes of it.. The technology is being used for racial profiling and surveillance of the Uighur Muslim population, to the outrage of the international community.
China also uses facial recognition to identify pedestrians crossing off-road… and fine them, or to identify students at the entrance to the institutions or to monitor their expressions during classes to check their attention.
Since the end of 2019, facial recognition is also mandatory for Chinese who wish to order a new SIM card or take the metro. The real objective of the Chinese government is to build a database of the faces of all citizens so they can be identified at any time via video surveillance.
Chinese researchers have even created a 500 MP camera capable of identifying all the spectators in a stadium instantly. In Hong Kong, the government prohibits pro-democracy demonstrators from wearing masks to prevent them from escaping facial recognition…
In Russia, in Moscow, surveillance cameras scan the streets using facial recognition. Coming soon, police officers will be equipped with goggles with facial recognition cameras as is already the case in China as well.
In the United Kingdom, the Metropolitan and South Wales Police have already tested facial recognition to detect individuals among spectators at football or rugby matches, on the streets of major cities or during music festivals and commemorations. Facial recognition will also be used extensively at the Tokyo 2020 Olympic Gamesfor surveillance and security purposes.
Facial recognition has several advantages. First of all, this technology allows to increase the level of security in society when coupled with CCTV. As cities become denser and more densely populated, this innovation makes it possible to maintain a level of security that peacekeepers can no longer provide on their own.
It also allows to Simplify the authentication process on electronic devices such as smartphones, while increasing the level of security over other methods such as passwords or fingerprints. This is undoubtedly the main tangible benefit for the average consumer…
Unfortunately, facial recognition also raises many fears and concerns about security and privacy. If Unfortunately, a hacker can access the collected data… by this technology, they could be exploited for malicious purposes.
Authorities and private companies could also use it to track individuals. Just as geolocation data or information about our activity on social networks already exist, facial recognition data could be used to build a “profile” for each of us. It then becomes possible for governments and marketers to “predict” our behaviour for the purposes of crime prevention or advertising targeting .
Of course, it is legitimate for citizens to be concerned about facial recognition being used for video surveillance purposes, especially since it is sometimes used without the citizens’ knowledge, for example in the King Cross area of London. In the near future, you may no longer be able to walk the streets without being identified at all times. by cameras. So this technology in itself is a threat to privacy and confidentiality.
Facial recognition can mistake you for a criminal.
Moreover, the problem is compounded by the fact that Facial recognition is still unreliable. Even though the performance of the systems has increased 20-fold between 2014 and 2018, and the error rate has fallen from 4% to 0.2%, there is no such thing as zero risk.
Images from real world cameras may be blurred or poorly lit, faces may be covered, and individuals may have aged relative to the reference photo in databases. Indeed, one person can be confused with another person.. So it’s possible that you’re wrongly accused of someone else’s crime…
Facial recognition is skewed
Also arises the problem of biases that granulate algorithms…. Because neural networks are trained mainly on photos of white males, they are less effective in identifying women or people of colour. A study by Colorado researchers also shows that mainstream systems are unable to correctly categorize trans or non-binary people.
In 2018, in a test conducted by the ACLU, the ACLU discovered that Amazon Rekognition software confused 28 members of the US Congress with criminals. The people involved were mainly African-American or Latin Americans. This lack of precision with regard to minorities amplifies the risk that innocent people will be wrongly accused…
Huge power in the hands of the GAFAM
Another problem is related to the fact that facial recognition is developed by tech giants such as Google, Facebook, Apple, Microsoft and Amazon. Governments and authorities have no choice but to turn to the AFSM… to take advantage of this innovation.
As a matter of fact, this technology only reinforces the already exaggerated power of these colossal private enterprises. In September 2019, Amazon announced it would create its own law to regulate the use of facial recognition…
Facial recognition surveillance is expensive, very expensive.
In addition to raising ethical issues, facial recognition poses another major problem: its widespread use is extremely expensive. Cameras generate huge volumes of video, and the resources needed to process and analyze this video are astronomical.
In addition, given the time required to review and analyze all of these videos, a criminal will have plenty of time to get away… after being caught on camera. In other words, facial recognition is currently far from being a viable solution for ensuring security in our societies…
Should we ban facial recognition?
In July 2019, the city of San Francisco has decided to ban completely the use of facial recognition by the government and authorities in order to protect the privacy of its inhabitants. It is the first American city to make this decision.
She was then followed by the cities of Oakland, Berkeley and Sommerville.. Several Democratic candidates in the presidential elections also raised their voices against this technology. One of them is Bernie Sanders, who is calling for a federal ban on its use by the police.
In an article dated 17 October 2019, the New York Times calls for going further by banning public recognition for both the public sector and private companies. In the eyes of the prestigious newspaper’s journalists, it is important to ban this technology before becoming dependent on it. to the point of accepting its dangers as a necessity.
As long as the use of facial recognition is not properly supervised and regulated, it will be impossible to avoid abuses and abuses. This is the reason why the only solution seems to be to ban this technology outright….
However, many industrialists and legislators doubt the need to ban this technology. In their view, its benefits are more numerous and important than its harms. They also believe that mistrust of facial recognition is a common reaction to new technologies, which was also observed when fingerprint sensors first appeared. The debate remains open on whether or not to ban facial recognition.and is likely to give rise to considerable tension between its supporters and detractors in the years to come as its adoption becomes more widespread .
Facial recognition in France
In France, as in the whole of Europe, the use of facial recognition is regulated by the RGPD, as are all technologies related to data collection. However, it is already used by many businesses and private companies for video surveillance purposeswhich is of great concern to the CNIL. The French data protection authority had also sounded the alarm when the city of Nice decided to test this technology to ensure security during its annual carnival.
From November 2019, the government will also turn to facial recognition with the launch of the Alicem mobile application. This will make it possible to connect to the various French public services using a selfierather than having to log in with a password. Although use of this application will remain optional, many fear that it is only a first experimental step before widespread, or even imposed, use in the hex. The association la Quadrature du Net believes in particular that this application goes against the RPGD …