Whether it’s surveillance cameras in public spaces, automated border control, or simply unlocking your iPhone, artificial intelligence-powered facial recognition has taken root in our lives. Not everyone accepts this without worry. The feeling of permanent surveillance and the misuse of data that reveals the behavior and whereabouts of fellow citizens leaves a queasy feeling. In countries with repressive regimes like China, facial recognition is already being used against their own citizens.
While these concerns are valid, I also want to highlight the positive use cases, because as always, every technology has its two sides. In this case, how facial recognition can help to reappraise history and give victims a name and a story.
American Civil War
The American Civil War Photo Sleuth project uses facial recognition software to identify soldiers from the American War of Secession, which took place between 1861 and 1865 and cost the lives of nearly 600,000 soldiers, in old photographs. This involves the collaboration of computer scientist Kurt Luther, of Virginia Tech University, and Ron Coddington, editor of the Military Images photo archive, who launched the project in 2018. The starting point was an exhibition on the American Civil War, where Luther had come across a photo of his great-great-great-uncle.
Of the estimated 40 million photos of soldiers taken with the young technology during the Civil War, only a tenth exist today. But the names of the people pictured are hardly known, their identities not preserved. And that’s exactly what this project aims to change. Members registered on the website can upload photos and add biographical data. The facial recognition software then attempts to match photos. Already in the first month, 88 people could be identified, according to the report in Slate magazine.
Even though the resolution of the photos doesn’t always quite allow for identification with today’s technologies, the project organizers are aiming to identify 100 percent of the people in the pictures.
The American project has inspired others as well. Google computer scientist Daniel Patt and financier Jed Limmer are using specially developed facial recognition software to analyze the portraits of 177,000 people in 35,000 photos held by the United States Holocaust Memorial Museum. In From Numbers To Names, the aim is to give back the names and thus dignity to those murdered by the Nazis.
Once again, registered users can participate in the analysis by uploading their own photos and adding biographical details. In addition, the 1,256 hours of film reels held by the museum will also be analyzed.
We easily overlook when looking at these black and white pictures that the people depicted on them had names, family, history and their dreams and hopes. Just as often the simple coloring of black and white images brings us abruptly closer to the past, the faces of the emaciated and abused figures seem familiar when we have a name to go with the person.
Here, too, the project strives to give the nameless figures back their identity and thus their dignity. Many descendants of those murdered in Nazi concentration camps still do not know when and where exactly they were killed. With the help of this project, many uncertainties could be solved.