Faces in the Crowd
This has been a year of super-recognition.
Posted December 14, 2020
Can you still remember the face of the person who sold you your first iPhone way back when? If you answered yes, then you might indeed be a so-called super-recognizer. Think you’ve got what it takes? Take the online test to find out.
James Dunn, a cognitive psychologist at the University of New South Wales who’s studied the results of the test above since 2017, told me that super-recognizers inherit their rare superpowers — it’s not a skill you can train. (Worldwide, only one percent of the population qualifies; on the other side of the spectrum, about two percent suffer from so-called Prosopagnosia, face blindness, that is, they really have a hard time recognizing other people’s faces.)
Dunn says that super-recognizers often prefer not to disclose their innate abilities so as to not scare off people in social settings or relationships. He shares the true story of a super-recognizer who recalled someone in a park in Paris taking photographs whom he had met briefly ten years earlier. This raises another question: Is it ethical for super-recognizers not to reveal their unusual abilities?
This is a particularly relevant question when you consider that in some aspects human super-recognizers are superior to AI-based facial recognition. According to Dunn, AI can be quite accurate under controlled conditions, like when someone is standing in a specific spot at the airport and looking straight into a camera. But when conditions are less controlled, lighting is bad, or only a part of the face (for instance, in face mask times) can be seen, human super-recognizers perform better.
That said, AI-based facial recognition is addressing those deficiencies and increasingly focusing on the eyes, as this new study claims. (Interesting detail: Red or black face masks apparently confuse the algorithms more than blue or white ones.)
Seeing and being seen on Zoom
This year we had to learn to navigate most parts of our social life with most parts of our face covered. Our tears were visible, but not our smiles. That pretty much sums up 2020.
“A mask tells us more than a face,” Oscar Wilde said. This year masked and unmasked us, and it is fitting that the title character of The Mandalorian, the masked bounty hunter in Disney’s Star Wars sequel who only shows his real face once, in the season two finale, became this year’s most popular character in storytelling.
In real life, because of social distancing and face masks, we saw fewer (of the) faces of others. Online, however, we saw more of our own. Watching ourselves speak and listen incessantly on Zoom forced us to study our little mannerisms and ticks as our gaze was thrown back at us. We watched our futile attempts to relate, so connected and so lonely, so close and so far away at once.
There is no intimacy without vulnerability, and on Zoom, we can’t really get hurt (unless we do something no one is supposed to see). It is essentially one big mask we are wearing, and the only revelation on Zoom is that it makes our wearing masks at work so obvious. And yet, from time to time, we are able to find communion in this faux intimacy, and it is still better than real isolation.
We’ve developed a whole new appreciation for the importance of the human face this year. Faces are our faces, they are our faces, the windows of our soul, the most visible manifestation of our emotions, the only features we can truly trust, in others and ourselves. We become vulnerable when we take our masks off, because we can be seen or seen through, recognized or exposed. We become human. “Enemies are people whose story you haven’t heard, or whose face you haven’t seen,” holocaust survivor and peace activist Irene Butter said.
At The Great Wave festival I co-hosted this year, the Spanish performance artist and facilitator Esther Blázquez-Blanco hosted a session, “I Came Here to Speak About Love,” in which she simply stared into her webcam for an hour, silently and supported by live music, and invited attendees to be with her, following closely the subtle intricacies of her face.
Faces that aren’t just masked, but are also entirely missing from public life
The painter, writer, and teacher Riva Lehrer describes the human face “as a theater that performs the actor inside, in flickers and puckers and pulls of 42 tiny muscles, in the rise and fall of blood that swirls with our emotions.” A disabled person herself (she was born with spina bifida and had scores of surgeries ever since), she paints portraits of people who, like her, have invoked their physical identity to challenge whatever it is we call “normal.” In an article about how the pandemic affected her work, she writes, “many of my collaborators — they are not mere subjects, but partners in creation — are disabled, or queer, or trans, or people of color; those, in fact, who are most at risk from Covid-19. Faces that aren’t just masked, but are also entirely missing from public life.”
This year, as we struggled to see the faces of loved ones and to confront our own, it was these invisible faces we began to see: those who usually remain unseen or operate in the shadows, in the economic margins of our society, undervalued and undercompensated, from parents to teachers, from artists to nurses, waiters, garbage collectors, and sex workers.
This year has been a year of recognition: seeing things we hadn’t seen before and understanding how we see.
Lehrer’s job as a portrait artist got disrupted by the pandemic, like so many others. “The virus has stolen your face from me,” she writes. “All I can do now is reconstruct the mysteries of who you are under the mask.” On Zoom calls, she tries to recognize the symbols and stories written in a face that are the fragments of the complexity of a human life. She finds solace in the fact that we are all portrait artists now, trying to “reassemble one another from fragments, from pictures on our screens and glimpses under the mask.”
Facial recognition, deep fakes, and the heart
Faces are tokens of our identity. Being seen is empowering; it is humanizing. Recognizing our faces at face value, on the other hand, is an act of power and potentially dehumanizing.
This is why facial recognition technology is so delicate. It’s been programmed with embedded bias and used to identify and classify millions of people without their consent. And it remains a highly controversial issue outside and inside the tech industry, as the recent debate over Google’s pushing out one of its leading AI ethicists, Timnit Gebru, shows.
How we see things is shaped by our beliefs, but seeing is no longer believing, as the rise of deep fake videos, made possible by machine learning, has shown us. Replacing an image or video with someone else’s likeness, deep fakes have gotten so proficient that it is increasingly hard to tell the difference between the manipulated content and the original source. Microsoft and other firms have developed AI that can detect deep fakes, but the use of these tools is not yet widespread, and technical challenges remain. And AI may be the wrong confidante anyway for telling us what’s true and what’s not.
I can’t help but think of Antoine de Saint-Exupery’s Little Prince and its most famous line: “It is only with the heart that one can see rightly; what is essential is invisible to the eye.” It is a line that has been quoted all too often, found its way onto millions of postcards and other trivia, and has been effectively rendered a cliché. But, like so many other seemingly trivial things, this year has reminded us of how true it is.
When it comes to recognition, we humans must continue to work on the one superpower that distinguishes us from machines: to see what only we can see.