Business school professors trained an AI to judge workers' personalities based on their faces.
Pluralistic: Business school professors trained an AI to judge workers’ personalities based on their faces (17 Feb 2025) – Pluralistic: Daily links from Cory Doctorow
Ahh, phrenology is apparently scientific if you do it with AI?
I fucking hate this timeline
It was "scientific" when they'd "confirm" it with stats.
What they call AI today is a family of obscure statistical instruments pretending to carry truth in that trait alone.
No, other than having stats you should also know and be capable of proving how those stats apply to the task at hand.
And they use the all-powerful electronic computation machine as a piece of technomagic to give it credibility.
Have you read Klemperer's book on Third Reich's language? I recommend it highly. Nazis used a lot of names for their policies, the subtle semantics of which are usually lost when translating from German. They used terms from radio and from automobile industry, for example.
Those who produce MBAs at it again.
I used to joke about, but there really is MBA-> fascism pipeline going on
When your education revolves around dehumanizing people and turning them into abstract numbers, it's not that far of a leap, unfortunately.
I don't know which machine learning textbook it was, but in the first few pages of it, the author warns about the stupidity and dangers of this exact same thing 🤷♂️
We are one step closer to building the AI that can determine which one is cuter: a specific photo of a kitten or one of a puppy. Just imagine what you could do with such technology!
Parasites always trying to justify their own existence
Great article, you should x-post to !aboringdystopia@lemmy.world
Yeah what could possibly go wrong?
_
Shitty title. Business school professors claim they trained an AI to judge workers' personalities based on their faces.
The article talks about why that is such a stupid and terrible proposal.
They've been doing this with "natural intelligence" for ages.
Makes sense. How physiognomy was used in the 19. century was bullshit (phrenology with skull measuring and whatnot) but the face is a display of your genetical base, your hormonal exposure while growing up and of your health currently. Btw, that's why the face is important in dating.
And you do use it the same way in social context, although subsconsciously.
but the face is a display of your genetical base, your hormonal exposure the last few years and of your health currently. Btw, that’s why the face is important in dating.
Everything is a display of everything else affecting it.
You are in some sense correct.
But using statistical instruments requires deep understanding of how they work. The article hints at that too.
And you do use it the same way in social context, although subsconsciously.
My experience is very different. When I see faces on photos, I get a completely different impression than seeing their owners personally and talking to them. Including romantic context.
Also what’s attractive to you might not be for me.
Sure, you don't see traumata etc on the face. It's just a base estimate and some people are better in it than others.
Its also affected by which direction the sun hits your house from, your commute to work (if applicable) and what kind of diet you had when growing up.
Look, i'm an Asperger and use the same mechanism regularly, to guess how the person would react. It's not obscure and not bullshit. People just don't like that they are more transparent than they think.
Makes sense.
Sure if you know nothing about biology, sociology, "AI", etc.
I do. Seems like you don't know enough.
This just one scary side of AI.
The idea of corporate level integration of this stuff is straight out of Black Mirror.
We're right around the corner from the corpo AI keeping tabs on your pupil dilation as you read your emails. If we aren't there already.
Just reading the title, sounded like a great way to automate racism.
And then I read:
So uh yea, doesn't seem like a great idea.