MEPs on the European Parliament’s Civil Liberties Committee have questioned EU plans to establish a bloc-wide facial recognition database for use by police authorities, citing the potential abuse of data as well as the likelihood of false positives.
As part of a planned extension of the EU’s 2008 Prum Decision, which allows for the exchange of DNA, fingerprint and vehicle registration data, member states have proposed that police authorities be given powers that permit them to share facial images.
Under the so-called ‘next-generation Prum’ plans, MEPs were told by experts on Tuesday (22 September), that there could be certain challenges to privacy rights that arise as part of the introduction of new data categories into the agreement, such as facial images.
Dr Niovi Vavoula, from Queen Mary University of London, told MEPs, the risk of ‘false positives’ and the possibility of ethnic minorities being unfairly targeted due to algorithmic bias are among the concerns.
This was a particular point also raised by Chloé Berthélémy of EU digital rights group EDRi, who said that it was “appalling” to consider expanding Prum to cover facial images in light of the concerns related to false positives and the “systemic racial bias that our law enforcement and criminal justice departments demonstrate”.
Dr Rafaela Granja, of the University of Minho, referred to recent research that surveyed national contact points (NCPs) across EU countries, which are responsible for facilitating data exchanges.
The study found that based on NCP responses, issues related to “false positives, lacking standards for procedures and issues of accountability and transparency,” are foreseen with the extension of the Prum framework.
MEPs are apprehensive about the extension idea.
“Even in the last mandate, it was said that the quality of data is of utmost importance for all these things,” S&D lawmaker Birgit Sippel said. “And still we are fighting with it, we are talking about false matches and about false positives as if this were normal.”
“But these false positives have consequences for investigations, investigations might go into a wrong direction,” she added.
Renew Europe’s Sophie in’t Veld questioned what approach the EU should take with regards to dealing with abuses of power that may occur as a result of certain nations using new provisions set out in the ‘next generation Prum’ which could be used to capture facial images of political opponents or protesters.
The European Parliament only plays a consultative role on the Prum agreement, so the Greens’ Sergey Lagodinsky suggested that a deeper fact-based assessment of the potential ramifications of the new system should take place.
“I think that the overall conclusion is very clear: evaluate before expanding,” he said. “We have to have a proper fact-based evaluation of the system so far, before we move on, to expand or change it.”
Despite the concerns raised, however, there were some lawmakers who were more positive about the notion of the Prum decision being expanded to include new datasets.
The EPP’s Tomas Tobé noted that a ‘high exchange’ of data between EU police forces is necessary for the fight against crime, but that there should be a better understanding on the usefulness of facial recognition data in this context.
Council conclusions dating back to 2018 suggested the treaty expansion and invited member state experts – as part of the Council’s Working Party on Information Exchange and Data Protection – “to evaluate the Prum workflow for further developments with a view to possible new biometric technologies, e.g. face recognition systems.”
Reports suggest that in order to judge the feasibility of incorporating facial recognition technologies into the Prum agreement, the Commission contracted international consultancy Deloitte to conduct an assessment, to the sum of €700,000.
The study is currently being discussed in Council working groups, with a public update on the plans due in October.
[Edited by Sam Morgan]