The senators have just adopted a bill authorizing the experimentation of facial recognition in the public space. A controversial measure, which raises fears of a generalization of the law and possible abuses.
Little by little, video surveillance has become part of our daily lives, for better or for worse. While video surveillance systems initially relied on analog devices (cameras, video recorders, etc.) and human operators, the development of automated digital technologies is gradually taking over. If they are more efficient, they are also more worrying, and it is not going to improve! On June 12, the Senate adopted in first reading a bill aimed at creating a legal framework for the use of biometric technologies and which authorizes, on an experimental basis, facial recognition in public spaces, for a period of three years. Tabled by senators Marc-Philippe Daubresse (Les Républicains) and Arnaud de Belenet (Union Centriste), it was voted 226 votes in favor against 117. A text that is of great concern and must now pass through the hands of the National Assembly.
Facial recognition: three years of experimentation
As a reminder, facial recognition is a biometric recognition technology whose purpose is to identify a person based on their physical or biological characteristics. It is therefore a particularly sensitive and controversial technology, particularly in terms of data protection and the abuses it can easily cause. The text adopted by the Senate lays down red lines that it must above all not cross, by prohibiting “any categorization and notation” people “on the basis of their biometric data”. This includes : “prohibition of social rating, prohibition of the categorization of individuals according to ethnic origin, gender or sexual orientation except in the case of scientific research, prohibition of the analysis of emotions except for health or research purposes, prohibition of real-time remote biometric monitoring except very limited exceptions with considerable safeguards”as reported Public Senate.
The text distinguishes between the use of a posteriori biometric recognition, that is to say, in the case of facial recognition, on images collected and then processed, and biometric recognition in real time. It indicates the cases where facial recognition could be used, namely in the investigation of “the most serious crimes” and, in the context of intelligence and security, “to the fight against the risk of terrorism”. Thus, a posteriori facial recognition will be authorized in the context of investigations and investigations relating to exceptionally serious offences, namely terrorism, arms trafficking, attacks on persons punishable by at least five years in prison or procedures for finding missing or fugitives. In this context, investigators will be able to use this technology to identify people listed in the criminal history file (TAJ) and to use video surveillance images already collected.
As for real-time recognition – which is much more sensitive – it will only concern the most extreme cases, namely terrorism, attacks on the fundamental interests of the nation and offenses relating to organized crime. In any case, the authorizations issued – either by the Prime Minister, or by the prosecutor or the examining magistrate – will only be valid for forty-eight hours. These measures must be tested for three years and will be subject to “robust systems of prior authorization and permanent control”.
Facial recognition: a text that deeply divides
This bill greatly worries the senatorial left. As indicated The worldthe socialist Jérôme Durain warned against “technological solutionism” and judged that he “no point in rushing” on this topic “ethics”. For her part, the president of the CRCE group – with a communist majority – Eliane Assassi regretted that “what yesterday was dystopia is materializing step by step before our eyes, without any public debate”. As for the ecologist Thomas Dossus, he denounced the “heavy dose of hypocrisy” from a text by “fake fanatics”. He still managed to add an amendment according to which “this experiment can in no way prejudge the sustainability of these treatments.”
The bill must now pass through the hands of the National Assembly, where it risks being less well received. Because if the text passed in the Senate, it is thanks to an alliance between the right (The Republicans) and the centrist Union. The Renaissance group (formerly La République en Marche), under the name of the Rally of Democrats, Progressives and Independents, voted against. We can therefore expect that the deputies of Renaissance, but also of the NUPES, oppose this text and obtain a majority.
Facial recognition: towards a generalization of the law?
The time to pass such a law is not necessarily the best one, given that the European Union is in the midst of a debate on the subject. Also, by adopting such a text, France risks coming into conflict with European law, which should arrive by the end of the year. And based on initial reports, it appears that the European Parliament is leaning more towards a permanent ban on the use of automated analysis, human characteristics, and other biometric and behavioral signals.
Moreover, despite the amendment, there is a high risk of deviations, the devices announced as experimental and provisional tending to become permanent in terms of security. Propose an experiment is actually a way to accustom the population to better pass the pill. Moreover, even if the bill is limited for the moment to extreme situations, its use can be extended later. We have just had a perfect example of this with the extension of night searches in the case of the Justice Law – initially reserved for cases of terrorism and organized crime, they now extend to crimes of flagrante against persons.
Video surveillance and AI: a dangerously explosive cocktail
Artificial intelligence is increasingly integrated into video surveillance. On June 8, a law was passed in the National Assembly, authorizing the use of AI for augmented video surveillance – called “intelligent”, which analyzes captured images in real time – during the Olympic Games. of 2024. Article 7 in particular sows discord by approving the installation of experimental devices of automated video surveillance“which will use artificial intelligence (AI) algorithms capable of detecting situations dangerous to the safety of people, such as crowd movements”, but also abandoned packages, “abnormalities” and suspicious movements (see our article).
A priori, this therefore has nothing to do with biometric data and facial recognition. However, some deputies had considered that the analysis of the behavior and movements of people already corresponded to the processing of biometric data, since the algorithm will make it possible to recognize people, without necessarily identifying them. However, recognition means providing a description detailed enough to enable agents in the field to locate a person. A decision which is causing debate and which even worries the National Commission for Computing and Liberties (CNIL).