The EU wants to use biometric data, drones, and self-learning systems with artificial intelligence (AI) to create a comprehensive surveillance network against migration. At a meeting between the EU border agency Frontex and lobbyists, companies from the surveillance industry were present alongside weapons manufacturers and other representatives.
EU’s New Pact on Migration and Asylum adopted last year, which provides for Frontex to be expanded to a large extent, including its technology, served as a background of the meeting. The Union has discovered a testing ground for the practical implementation of novel surveillance methods, particularly when it comes to dealing with migration at the external borders: Frontex, for example, has been relying on systems controlled by algorithms for some time. The aim is to replace human decision-makers or to provide them with technology-supported assistance.
Migrants are monitored even before they arrive. Frontex has been cooperating with the Israeli company Windward for years. The company, owned by former CIA director David Petraeus among others, is tracking boats in the Mediterranean by combining position and weather data. With further data from shipping companies, owners of ships and satellite images, ships that are identified as suspicious are marked by the agency staff.
The basis for doing the calculation is, on the one hand, the register of the International Maritime Organization; on the other hand, the company claims it is using its own data about 400,000 boats. Furthermore, freely available data from the internet are used to compare, and the information is processed together with other software. This is how the agency sends reports to the EU Commission, the EU Council and other border authorities. The Eurosur surveillance network, set up by Frontex, is used for this purpose.
Biometrics and drones
In the future, large drones and unmanned Zeppelins will also support Frontex in its surveillance of the Mediterranean. Moreover, the agency’s aircraft are to be equipped with a wide variety of sensors to detect migrants when they are travelling in environments that make search operations difficult – in forests, for example, but also at night or in adverse weather conditions. The Austrian Institute of Technology (AIT) in Vienna is also working on the project, which is called “Foldout”.
Frontex will also push ahead with the use of biometric data: from 2023, every person from a third country who wants to cross an EU external border will have to leave their fingerprints and an image of their face. With this, the EU agency is increasingly focusing on the so-called “predictive” work – it wants to use the systems to identify migration flows before they even occur.
Before purchasing the relevant hardware and software, the agency met with European companies Airbus, Indra, Leonardo, GMV, as well as Japanese biometrics company NEC, according to the leaked papers.
Controversial ‘lie detectors’
In addition, starting in 2023, even before they cross a border, foreigners will be required to disclose their itinerary to digital “avatars.” Using automated systems, the agency’s Frontex staff will then be notified of any behaviour identified as suspicious. This is done by analysing facial expressions. The effectiveness of this kind of “lie detectors” is, however, controversial. Nevertheless, facial expressions and behavioural analyses have been in use at Frontex for years.
Currently, a lawsuit by MEP Patrick Breyer (of the German Piratenpartei) is pending against the – according to the EU – AI-based “video lie detector” iBorderCtrl. This is supposed to detect whether people are lying when answering to questions based on facial expressions. The system has been trialled at external borders.
The proceedings before the ECJ are still ongoing. Breyer wants to know above all how the software was ethically assessed, on what legal basis it is based, how it is advertised and what were the results during the testing period. “The European Union keeps funding illegal technology that violates fundamental rights and is unethical,” Breyer criticizes. From his point of view, it is a “pseudo-scientific” development that should be disclosed on the basis of transparency complaint.
Artificial intelligence is not neutral
The lawyer Petra Molnar, who together with the civil rights organisation Edri, among others, has published a report on surveillance at the borders, criticises the EU agency: “Frontex has long positioned itself as the leading institution for the surveillance of Europe’s borders,” she tells the Der Standard. The latest developments would show the “profound curtailment” in human rights faced by those who are subjects of the “cutting edge technological developments” anyway.
AI would never be neutral, she finds. This has to do with the fact that, usually, information from the past is used to train such systems. If there were unfair decisions, they would be continued and reinforced further, Molnar explains. “Racism and discrimination is perpetuated by biased data sets,” the lawyer says. “The hubris of ‘big tech’ and the lure of quick fixes ignore the systemic problems because of which groups are forced to migrate in the first place,” she criticizes.