Law enforcement authorities in 11 European countries are already using biometric recognition systems in their investigations. Eight more are to follow, a new study found, warning of the technology’s impact on fundamental rights.
Police in Austria, Finland, France, Germany, Greece, Hungary, Italy, Latvia, Lithuania, Slovenia, and the Netherlands employ facial recognition technologies for ‘ex-post identification’ in their criminal investigations. Croatia, Cyprus, Czechia, Estonia, Portugal, Romania, Spain, and Sweden are expected to follow suit soon.
Ex-post identification is where footage is checked after an incident, not in real-time.
“The distinction between “real-time” and “ex-post” is irrelevant when it comes to the impact of these technologies on fundamental rights. Ex-post identification carries in fact a higher potential of harm, as more data can be pooled from different sources to proceed to the identification,” said Francesco Ragazzi, associate professor at Leiden University and author of the study.
The study, published on Monday (25 October), was commissioned by the Green group in the European Parliament, a vocal proponent of a total ban on biometric recognition systems in publicly accessible spaces.
The most developed form of biometric identification is facial recognition software which matches one captured image to another image stored in a database.
The report warns that “there seems to be little understanding of the ways in which this technology might be applied and the potential impact of such a broad range of applications on the fundamental rights of European citizens.”
The study focuses on so-called non-cooperative searches, where the system tries to identify a person without their consent. It notes that the deployment of these technologies is still limited in scope and scale across Europe, ranging from the identification of individuals to mass surveillance.
In cooperative searches, for instance, facial recognition to unlock smartphones is not considered to currently pose a risk of mass surveillance. However, the report notes that the situation might change if the legal framework were modified, as these cooperative systems have accumulated massive amounts of personal data.
Two developments are pointed out as making mass surveillance more likely: the expansion of biometric databases and the piloting of several systems connected with biometric identification software.
“What we see with these projects is that they increasingly follow a ‘fait accompli’ strategy. They are generally presented as a pilot project, asking for some special circumstances about the deployment of the technology. They ask for permission later,” Ragazzi added.
In 2017, the Brussels airport deployed four facial recognition cameras without notifying the relevant supervisory authority. A neighbourhood in Rotterdam launched a “burglary free” project to detect suspicious behaviour using smart street lamps.
Nice started trailing facial recognition technology in streets. The use of biometric identification tools in high schools was also tested but consequently declared illegal. By contrast, facial recognition tools are now used to identify pupils in the school canteens in the UK.
In Germany, Berlin, Hamburg and Mannheim have all deployed facial recognition software to test their capacity to detect suspicious behaviour.
“Justification as a test approach is often used in Germany to argue for a deviation from existing rules and societal expectations,” the report reads.
The study notes that these pilots tend to start in a legal grey zone and, if left unchecked, might have the long-term effect of normalising surveillance. In particular, monitoring suspicious behaviour might have a chilling effect on individual freedoms.
Furthermore, in most cases, the infrastructure such as cameras and microphones has been disabled, but they remain in place.
Earlier this month, the European Parliament adopted a resolution calling for strict rules on employing AI systems in law enforcement, notably pushing for a ban for facial recognition technologies in public spaces.
The report had strong support from the liberal, social-democrats, leftist and green groups, while it almost equally split parties from the right and extreme right. By contrast, the Christian democrats voted almost unanimously against the ban.
“While there need to be strict safeguards and certain prohibitions, a complete ban of facial recognition disregards the benefits such technology can have, for example when tracking down criminals or even as part of other use cases such as training autonomous vehicles for recognising humans,” said Axel Voss, an influential MEP in the Christian-democrat camp.
The ban might be integrated into the Artificial Intelligence Act, which Brando Benifei, rapporteur in the internal market and consumer protection committee (IMCO), called for. However, other parliamentary committees have disputed IMCO’s leadership, opening a dispute over competencies.
To solve the dispute, the chair of the Conference of Committee Chairs, Christian democrat Antonio Tajani, recommended a joint leadership between IMCO and the legal affairs committee (JURI). In JURI, one of Tajani’s fellow party members is expected to take the lead, making a total ban less likely.
The final decision is up to the Conference of Presidents, which will discuss the matter on 18 November.
[Edited by Zoran Radosavljevic/Alice Taylor]