We’ve all met someone we instantly dislike – even before they open their mouths.
It’s not that you’re total strangers. A mutual friend is introducing you. But as the three of you stand there, there’s just something about him (or her) that ruffles your fur. Their thin smile? Their hearty handshake? Their ridiculous shoes?
No one ever said that being a human is fair. The fact that most of us are brutally unfair in how we assess strangers suggests that kindness has had to be beaten into us over many generations for us to wait for more than the blink of an eye to decide if we can trust them, let alone embrace them.
This idea of course leads to judging people by the colour of their skin, by their accent, or even by the shape of their heads. Phrenology was a 19th-century pseudoscience that believed people with differently-shaped heads had different personality traits.
I mention phrenology because at dinner with our next-door bubble-neighbours last night, we were discussing the Washington riots. I said it seems that facial recognition technology is advancing so quickly that the people analyzing your picture inside the Senate Chamber or at Pearson before your flight, or as you walk into Toronto General, can determine if your political leanings are liberal or conservative.
My friend (a doctor) scoffed: “Oh, that’s just phrenology.” He must have thought I’d turned anti-science. If anything, and tragically, the opposite appears to be true. Or at least coming to a truth near you.
Last week a Polish psychologist and Cambridge (and now Stanford) professor, Dr. Michal Kosinski, published a paper in Nature, one of the most rigorously peer-reviewed scientific journals in the world. The paper is titled “Facial recognition technology can expose political orientation from naturalistic facial images.”
Kosinski used a sample of 1,085,795 participants from the US, Britain and Canada and their self-reported political orientation, age, and gender. Their facial images (one per person) were obtained from their profiles on Facebook or a popular dating website. This was not a small sample.
He then created a facial recognition technology that he applied to the million-plus photos to determine if the faces of self-described liberals and conservatives differ consistently.
Kosinski’s algorithm ‘guessed’ political allegiance correctly 72% of the time. Remember, in this either-or world of liberal or conservative, this is “… remarkably better than chance (50%), human accuracy (55%), or one afforded by a 100-item personality questionnaire (66%).”
There was no appreciable difference between Americans, Brits and Canadians, nor between Facebook and dating websites, “and accuracy remained high (69%) even when controlling for age, gender and ethnicity.”
As Dr. Kosinski concluded: “Given the widespread use of facial recognition, our findings have critical implications for the protection of privacy and civil liberties.”
Let’s see. If you can tell someone’s politics from their face, you should also be able to make face-based judgements about their intelligence, sexual orientation, honesty, personality and even their violent tendencies.
And remember, these are very early days for facial recognition. Give it five or six years and the world of traditional highly-rigorous scientific enquiry around facial recognition, a world inhabited by Michal Kosinski, will be very much farther advanced than it is now.
The inevitable and scary question of course is, if his pioneering work proves out over time, how will any of us hide?
One of last summer’s big Netflix hits was The Trial of the Chicago Seven, written and directed by Aaron Sorkin. In the climactic court scene, the government prosecutor is grilling Abbie Hoffman, one of the defendants, played by Sacha Baron Cohen.
The prosecutor asks: “When you came to Chicago, were you hoping for a confrontation with the police?”
Prosecutor: “I’m concerned you have to think about it.”
Hoffman: “Give me a moment, would you, Frank? I’ve never been on trial for my thoughts before.”
That was then. It appears this is now.
8 thoughts on “The Plague-Ground – I see you. I know you. I jail you. I kill you.”
Ronan Farrow aces the phrenology test. Among the extremists storming the capital was Donovan Crowl, a fifty-year-old former marine. Farrow in his January 14, 2021 article in The New Yorker, “A Former Marine Stormed the Capitol as Part of a Far-Right Militia” reports why he succeeded in his attempt to speak with Crowl:
“He repeatedly denied that he harbored racist views or had made racist comments, but he said that he believed in phrenology, the discredited pseudoscience that infers intellectual ability from skull shape, and told me that he had responded to The New Yorker’s request for comment after examining pictures of my head.”
Oh wow, that is fantastic !
Good morning Bob. Thank you again. This was an incredibly important piece of reporting. I hope there will be a part two: what to do about it? Because if there is no answer to that, we are just left in a hopeless place.
Indeed, we are left in a hopeless place. Not sure where to turn. Maybe become a Replicant? Seriously, it is the beginning of the…
“…how will any of us hide?” One word: Facemasks. I’ve been wondering lately how babies born in the past couple of years will learn to identify moods if all they can see are eyes. Are the algorithms able to detect political affiliations through the masks we now wear?
Hmmmm…intriguing question Valerie. Let me try to find out.
Masks are not going to offer protection from facial recognition. According to Vox “there’s a global arms race right now to develop facial recognition software that can track people, even when we are wearing masks.” And how insidious this is becoming was reported in the last week or so – China’s Huawei is busy backtracking after filing for a patent to identify Uyghur faces. https://www.ctvnews.ca/world/china-s-huawei-backtracks-after-filing-for-patent-to-identify-uyghur-faces-1.5269680
Even worse, they filed it in July of 2018. No one was looking?