Remote biometric identification is a colonial boomerang

Every step you take. When I first took stock of face recognition in policing in Europe back in 2019, it was a novelty. Six years later, remote biometric identification is everywhere. Serbia and Montenegro use live face recognition in large cities. The United Kingdom is moving towards a general deployment of the technology. In the EU, live face recognition is forbidden for the moment, but the AI Act allows for it in not-so-limited circumstances, such as “the threat of a terrorist attack.”

Some might trust the police to use this infrastructure legally and responsibly. Others might remember that many police forces have a history of using biometric recognition first and asking for permission later, as happened in Sweden, France or Slovenia. In Berlin, lawmakers are pushing for building a database of biometrics for the police, contradicting the AI Act. Remote biometric identification is widely used in the EU and, despite or because of the AI Act, its usage will likely grow. What’s the appeal?

Why oh why? If remote surveillance provided security, we’d have ample evidence by now. But we don’t. Numerous studies failed to show an effect on crime or safety, except in car parks. If it was about public safety, we’d expect police forces to invest at least some resources in techniques that do work in that regard, such as community policing.

Perhaps do police personnel enjoy shiny new gadgets. But, in my experience of men and policemen in particular, they prefer motorized vehicles. Why invest millions in software when you could buy Lamborghinis?

Remote biometric identification considerably tips the balance between demonstrators and the government. If the police can identify all members of a group, they are not protected by the mass anymore. Perhaps the push for biometric surveillance is driven by a fear of the crowd? It might, but evidence from Indonesia, where live face recognition is widely deployed, suggests that it doesn’t help much. Last month, during a Gen-Z revolt, the police totally failed to ensure that people protest peacefully and, although thousands were arrested, the government had to make concessions.

Such explanations (remote biometric surveillance reduces crime and helps control crowds, or acquiring new gadgets is just all play and fun) don’t really lead anywhere.

British Raj. The most widely known history of biometrics starts in London and Paris where, around 1900, detectives began to collect mug shots (“le Bertillonage”) and fingerprints. This is true but misses an important point: These endeavors were restricted to criminals and suspects. The first attempt to submit an entire population to biometric identification took place in 1860s India.

William Herschel was a British administrator in Bengal. Like all his peers, he could not ascertain the identity of Indians. His racism prevented him from telling them apart visually, and his distrust prevented him from relying on traditional institutions, also in use in Europe at the time, such as having others vouch for the identity of someone. The solution, Herschel found, was in the fingertips. He had Indians appose a fingerprint on contracts, on registration deeds, on claims for pension and more. His system worked so well he called it “a miracle.” (To read more about the fascinating history of fingerprinting in India, read Chandak Sengoopta’s Imprint of the Raj.)

Boomerang. Poet and politician Aimé Césaire wrote in 1950 that colonial techniques come back to Europe as a “terrific boomerang.” Biometric identification is one of those. Not only did the London police’s adoption of fingerprinting suspects in the early 20th century build on experience from India. The current wave of biometric surveillance also answers the same needs William Herschel had 150 years ago.

Colonial governance is governance without trust. Violence, but mostly the destruction of existing institutions, breaks down all links of trust between authorities and population. Officials expect the population to constantly lie, and therefore mistrust everyone. They see the solution to regaining control in “schematizing, tabulating, indexing and recording everything in sight (and out of sight),” in the words of Edward Said. Remote biometric surveillance is not only in place to provide security (it doesn’t) or prevent gatherings (it can’t). It is the only means of control left when trust in the population has withered.

This is an excerpt from the Automated Society newsletter, a bi-weekly round up of news in automated decision-making in Europe. Subscribe here.