On December 8, 2023, European Union lawmakers were pleased to have reached agreement on the long-awaited proposed regulation on artificial intelligence (“AI Regulation”). The main parliamentarians then assured their colleagues that they had succeeded in including strong human rights protections in the text, in particular by excluding biometric mass surveillance (BMS).
However, despite the announcements made by European decision-makers at the time, the AI regulation will not prohibit the vast majority of dangerous practices linked to mass biometric surveillance. On the contrary, it defines, for the first time in the EU, conditions for the lawful use of these systems. MEPs and ministers from EU member states will vote on whether to accept the final deal in spring 2024.
The EU is making history – for the wrong reasons
The Reclaim Your Face coalition has long argued that SBM practices are error-prone and risky by design, and have no place in a democratic society. The police and public authorities already have a large amount of data on each of us; they don’t need to be able to constantly identify and profile us, objectifying our faces and bodies at the touch of a button.
Yet, despite a strong negotiating position from the European Parliament calling for a ban on most SBM practices, very little had survived the AI regulation negotiations. Under pressure from law enforcement representatives, Parliament was forced to accept particularly weak limitations around intrusive SBM practices.
One of the few safeguards that apparently survived the negotiations – a restriction on the use of facial recognition ex post [as opposed to use in real time] – has since been gutted in subsequent discussions so-called “technicals” which have been held in recent weeks.
Despite the promises of the Spanish representatives in charge of the negotiations, who swore that nothing substantial would change after December 8, this watering down of protections against facial recognition after the fact is a new disappointment in our fight against the surveillance society.
What is the content of the agreement?
From what we have seen of the final text, the IA Regulation is a missed opportunity to protect civil liberties. Our rights to participate in a demonstration, to access reproductive health care or even to sit on a bench could thus be threatened by omnipresent biometric surveillance of public spaces. The restrictions on the use of facial recognition in real time and a posteriori provided for by the AI law appear minimal and will apply neither to private companies nor to administrative authorities.
We are also disappointed to see that when it comes to “emotion recognition” and biometric categorization practices, only very limited use cases are prohibited in the final text, with huge gaps.
This means that the AI Regulation will allow many forms of emotion recognition – such as police use of AI systems to assess who is or is not telling the truth – although these systems have no credible scientific basis . If adopted in this form, the IA regulation will legitimize a practice which, throughout history, has been linked to eugenics.
The final text also plans to authorize the police to classify people filmed by video surveillance cameras according to their skin color. It is difficult to understand how this can be permitted given that European law normally prohibits discrimination. However, it seems that, when practiced by a machine, legislators consider such discrimination acceptable.
Only one positive thing emerged from the technical work carried out following the final negotiations in December: the agreement intended to limit public facial recognition ex post to cases relating to the prosecution of serious cross-border crimes. Although the Reclaim Your Face campaign called for even stricter rules in this area, this represented significant progress compared to the current situation, characterized by massive use of these practices by EU member states.
It seems that it is France which has led the offensive aimed at steamrolling our right to be protected against the abuse of our biometric data. As the Olympic and Paralympic Games approach in Paris this summer, France has fought to preserve or expand the powers of the state to eradicate our anonymity in public spaces and to use systems of opaque and unreliable artificial intelligence in an attempt to know what we think. The governments of other member states and the main negotiators in Parliament failed to counter her in this approach.
Under the AI Regulation, we will therefore all be guilty by default and put under algorithmic surveillance, the EU having given a blank check to mass biometric surveillance. EU countries will thus have carte blanche to strengthen surveillance of our faces and bodies, which will create a chilling global precedent.
This article is originally published on laquadrature.net