2023-06-18 14:02:00
If the whole planet is worried about the lack of regulation around artificial intelligence, the European Union is the first to react and to agree on a common position on a European regulation about AI. For its part, France is seeking to adopt the use of facial recognition AI for surveillance cameras now. A decision in total contradiction with the European provisions which will not be adopted before 2026.
You will also be interested
[EN VIDÉO] Humanity shares 70% of its facial expressions The approximately seven billion humans inhabiting the Earth would have another point in common. Finally, 16…
There is an air of contradiction between the positions of the European Union and those of France on artificial intelligence (AI). The subject concerns security and in particular the use of these AIs to improve it. This Monday, June 12, French senators voted on a proposal to loi to employ AI-assisted facial recognition as part of a three-year experiment.
A permanent use in real time on the public space which would supposedly be limited to sports and cultural activities, according to the senators carrying the proposal. The use of these tools was already voted on at the time of the so-called JO 2024 law, but certain amendments, in particular on real-time facial control, had been rejected. The senators therefore reoffend by wishing to reinject them today. While we have noticed the massive use of drones by the police the day after their application decree, it is certain that such an experiment should very quickly overflow its supposedly wise original intentions.
The argument of the senators to justify this job: the fight against terrorism. A strong argument, but even with the massive use of automated surveillance cameras monitoring all citizens in real time, nothing indicates that there would not be “holes in the racket” and that the risks would be reduced to nil. If there is little chance that this proposal will go much further, it nevertheless shows the techno-security spirit in which France is committed.
Recently, the Senate also said yes to a bill providing for turning on the microphones and cameras of smartphones in certain investigations. Similarly, the use of tools to encrypt communications (Signal, Telegram, Proton Mail) is an area that can generate some suspicion on the part of the authorities. This is what was revealed by the organization the Squaring the Net in a case dating from the end of 2020.
France champion of surveillance
But, as for this CCTV via IA, no offense to the senators, there is a catch. This Wednesday, June 14, the European Parliament meant that these AI-enhanced facial recognition systems were considered risky and should therefore be classified with the AIs whose use will be prohibited within the member countries of the European Union. Placed at the top of the risk scale, we therefore find these “real-time” remote facial recognition systems in public spaces. The same systems used “a posteriori” are also used with a few exceptions which make the application of the current French law JO 2024 debatable.
Also prohibit AI capable of biometrically categorizing sensitive data (gender, race, citizenship status, religion, political orientation), or even predictive police systems (profiling, location, previous criminal behavior). With their decisions, MEPs seek to ensure that AI can benefit from permanent human control, that it is safe, transparent, traceable, non-discriminatory and respectful of the environment. But this IA Act, ambitious in terms of ethics, will not happen tomorrow. The decrees applying the JO 2024 law and perhaps others, will have every opportunity to be taken before. And for good reason, for the final text of the European Parliament to be adopted, it will be necessary to wait until the end of the year.
Faster than legislative music
Its entry into force could even take place within at least three years, leaving the door open in the meantime to the unbridled use of these AIs dedicated to surveillance by the Member States, including France in the first place. And theIA Act European is not limited to this security side. Constraining the development of generative AIs like ChatGPT to certain ethics is part of the missions given to the European Union within the framework of this regulation. Here again, technology will go faster than legislative music, since by way of example and not to mention the American or Chinese high-tech bulldozers, during this edition of the Vivatech forum, the President of the Republic Emmanuel Macron announced an ambitious plan funded by France 2030, aimed at supporting players in artificial intelligence made in France. Not sure that the ethical supervision side is on the program…
1687155686
#France #tide #Europe