After several recommendations adopted in the last three years for the area of artificial intelligence (AI) and a market consultation, the EU executive presented, on 21 April, a proposed regulation to safeguard EU values and fundamental rights and user security, aiming to oblige systems considered as high-risk to comply with mandatory requirements related to their reliability.
"Non-compliance by an AI provider [...] may entail a fine of up to 6 percent of its global annual turnover," Margrethe Vestager, Executive Vice President of the European Commission with the portfolio of "a Europe fit for the Digital Age", explained at a press conference in Brussels.
Through a risk-based approach, the EU executive has proposed that in this first EU legislation for AI, tight controls should be imposed on systems considered as high risk and those of unacceptable risk should be banned.
However, although remote biometric identification systems are mentioned in the new legislation as "high risk", Margrethe Vestager clarified to the press that there are applications that are "not problematic", such as facial recognition or fingerprints, provided they are for personal use.
"One specific use of AI that falls both into the high-risk and the prohibition categories is biometric identification by remote identification, which can be used for many purposes and some of them are unproblematic, such as when it is used in one-to-one border control by customs authorities or when we are asked to sign with our fingerprints or by facial recognition," she specified.
On the contrary, "in our proposal, we focus on remote biometric identification, whereby many people are tracked at the same time", pointed out Margrethe Vestager, arguing that "any such use is highly risky from a fundamental rights point of view".
"And that is why we subject remote biometric identification to even stricter rules than other high-risk use cases," she justified.
He went on to say that "there are situations where [the rules for high-risk systems] may not be sufficient and that is when remote biometric identification is used in real time by law enforcement authorities in public places", so its use will now be prohibited in the EU as "there is no room for mass surveillance in European society".
The new law provides that all emotion recognition and biometric categorisation systems will be subject to specific transparency requirements, and should also be considered as high risk in cases of use in the areas of employment, education, law enforcement, migration and border control, for example.
Also according to Margrethe Vestager, "it will be up to member states to identify which national authority will be the best" to oversee the new rules, whether for example data protection or market surveillance.
In this dossier, which is a priority for the Portuguese presidency of the EU, Brussels then creates the first legislation aimed at AI, although those responsible for this technology are already subject to European legislation on fundamental rights, such as data protection, privacy and non-discrimination.