AI: From the repression of migrants to generalised surveillance

Fortress Europe wasn’t born yesterday. From the Roman limes to the Iron Curtain, the continent has taken pains to delineate its borders, to separate itself from the outside world, from the other. Borders change, are redrawn, and evolve with society. In the age of digital technology and artificial intelligence, some of these walls have become invisible, but they remain very real.

The border is near and far, everywhere and nowhere. Our newly developed and developing technologies allow us to project these demarcations into the immaterial, while algorithms and tools of mass surveillance and facial recognition make it possible to dilute, spread and stretch these borders, outwards and inwards.

In Solomon, Giorgos Christides, Katy Fallon, Deana Mrkaja, Marguerite Meyer, Florian Schmitz and Hibai Arbide Aza share the findings of a large collaborative investigation into how certain countries, including Greece, are making massive investments to stop migrants at their borders. “The arsenal at Europe’s disposal includes artificial intelligence (AI) systems, drones, thermal cameras, dialect detectors, data phone extraction and sophisticated surveillance networks,” the journalists explain. “Depending on the country using them, the aim of deploying these advanced and often costly systems is to help prevent migrant arrivals, scrutinize asylum claims and disrupt smuggling networks.”

As the popularity of mass surveillance technology grows, so do concerns about their pitfalls. “Critics counter that they are full of legal and moral pitfalls, undermining human rights, limiting access to asylum, infringing on migrants’ privacy, and can be used to facilitate collective expulsions—a practice that has been extensively documented and was recently described by the European Court of Human Rights as ‘systematic’.”

In some cases, methods designed to control migratory movement are also being used to control domestic populations, mobile or not. For Wired, in partnership with Lighthouse Reports, Caitlin Chandler reveals the scale of the market for “predictive travel surveillance” technologies. For several years now, companies have been promising the world’s governments programmes that use algorithms to track targeted individuals traveling by air: terrorists, traffickers, migrants. The aim is to smooth international traffic, facilitate the work of air traffic controllers and police forces, and ensure the safety of other travellers. “For anyone who travels internationally, these surveillance systems may provide some convenience,” explains Chandler, “but they can also flag you as a potential threat or even limit your freedom to travel, while giving you little ability to do anything about it.”

There is abundant potential for abuse: what if a user is incorrectly flagged as suspicious or, conversely, if a genuinely dangerous person is ruled out by the algorithm? There is also the issue of data retention and access. And then there are the societal consequences of a technology which, as one of the companies quoted in the article says, “enables a government to export its border to every single point on the globe where passengers can board flights, ships or trains bound for their territory.” The world is open. The border is everywhere. Decentralised.

The border between the repression of migration and repression as such is decidedly thin. With the rise of artificial intelligence in recent years, there has been a growing interest in assisted surveillance, and both industry and governments appear unmoved by the associated risks. This is especially the case in France: according to a collaborative investigation by Investigate Europe, Paris has campaigned to ensure that practices such as real-time facial recognition, the interpretation of emotions and the creation of files based on political and religious convictions are permitted under the European regulation on artificial intelligence, which comes into force in 2024.

“Imagine you are taking part in a climate demonstration, wearing a badge or holding up a placard”, write Maria Maggiore, Leïla Miñano and Harald Schumann, the authors of the investigation. “An ‘intelligent’ camera detects these signs, records your face and sends the images to the police for comparison with the file of people wanted for crimes committed by environmentalists. You don’t appear in that file, but the data is saved”. This is technology that the authors believe could be used to police migrants. “Imagine a castaway who has just landed on the island of Lampedusa. He is arrested and questioned using a camera capable of detecting emotions. The system registers signs of nervousness, fear and indecision: it concludes that he is lying about his age or origins. His asylum application is rejected”.

This is what France has obtained through insistent lobbying within the Brussels institutions. In the name of national security, the state will be able to search for individuals on the basis of personal criteria, including religious beliefs or political commitments. France, along with other European countries, has also called for prisons and border areas to be excluded from the definition of “public space”. “In a few weeks‘ time, Member States will be able to deploy emotional recognition systems on their doorsteps”, explain the journalists. This is despite the fact that artificial intelligence tools designed to recognise emotions have been widely criticised, not least because of the biases that plague their operation.

“This too is artificial intelligence”, says Félix Tréguer, founder and spokesperson for La Quadrature du Net, an association defending digital liberties. “The return of naturalistic theories, pseudoscience and arbitrary categories, now enshrined in powerful automated systems to implement state violence.”

But generalised surveillance is all the rage. Just look at the programme of Germany’s Christian Democratic Union (CDU). Just days before the February 2025 snap elections, they propose, among other things, extending the collection of telecommunications data, the use of facial recognition and the use of spyware by the state. As Constanze Kurz reports for Netzpolitik, “the vision that [emerges] is that of a future in which mass surveillance is ubiquitous.”

What the party also wants is to liberate this mass surveillance from any possible control. For the moment, such control does not exist, and in any case it seems difficult to apply, as Kurz explains: “This is CDU campaign noise,” she writes, “but the noise clearly indicates the direction the party will take if it accedes to the chancellery.”

This dual use of technology echoes the dual function that borders have always, historically, fulfilled: protecting populations from the outside world, while keeping those populations inside - and making sure they stay in line, if possible.

From the outside, Europe looks like a fortress. And from here on the inside, its walls may well resemble a prison.