Pulse of Truth

493 readers

71 users here now

Cyber Security news and links to cyber security stories that could make you go hmmm. The content is exactly as it is consumed through RSS feeds and wont be edited (except for the occasional encoding errors).

This community is automagically fed by an instance of Dittybopper.

founded 1 year ago

MODERATORS

krogoth@infosec.pub

-2

AI Can Tell Your Political Affiliation Just by Looking at Your Face, Researchers Find (gizmodo.com)

submitted 6 months ago by lemmydev2@infosec.pub to c/pulse_of_truth@infosec.pub

11 comments fedilink hide all child comments

A study recently published in the peer-reviewed American Psychologist journal claims that a combination of facial recognition and artificial intelligence technology can accurately assess a person’s political orientation by simply looking at that person’s blank, expressionless face. Read more...

all 12 comments

sorted by: hot top controversial new old

[–] glimse@lemmy.world 10 points 6 months ago (1 children)

I don't believe it.

[–] nahuse@sh.itjust.works 7 points 6 months ago (2 children)

Machine learning assisted phrenology!

History feels more and more cyclical every fucking day, I swear.

[–] MajorHavoc@programming.dev 3 points 6 months ago

Yeah. I actually didn't have "bring back Phrenology" on my "systemic regression into the dark ages" bingo card.

[–] shalafi@lemmy.world -3 points 6 months ago (1 children)

Bumps on one's skull is a far cry from facial features. We humans intake a wealth of data from faces.

[–] nahuse@sh.itjust.works 1 points 6 months ago

Facial expressions, sure. But this study measured still photos from volunteers.

[–] cheesorist@lemmy.world 6 points 6 months ago

just because ai does something doesn't mean it's done correctly.

[–] shalafi@lemmy.world 5 points 6 months ago

As a human I feel like I can make a damned good guess as to someone's politics by looking at them. We're really, really good at picking up clues from faces, even if we're not conscious as to why we're getting those clues.

Despite being an avid shooter, I'm very liberal. No one I've talked to, or been around, regarding guns has ever assumed I'm conservative. In fact, I've noticed they're damned careful to dance around politics around me.

Maybe it's the long hair? OTOH, I can be red necked out in my attire and holding an AR-15 and people still won't being up conservative views. And I'm in a very conservative area where it's safe to assume a guy that looks like me is a Trump voter.

I could see an AI correlating 10,000 facial cues to make an accurate guess. Interesting to think on.

[–] Umbrias@beehaw.org 4 points 6 months ago

"About as predictive as job interviews for job success"

So really fucking bad?

r = 0.23

Lmao.

Is this a joke? This isn't even close to a correlation. If anything this is evidence of a lack of correlation.

[–] alex_02@infosec.pub 2 points 6 months ago

I'm politically affiliated with that rich nasa furry that makes six figures.

[–] stevedidwhat_infosec@infosec.pub 2 points 6 months ago* (last edited 6 months ago)

"AI predicts your political affiliation based on the CURRENT stats." Fixed it.

This isn't intelligence. This is predictive intelligence based off a set of factors. This is how racism, sexism, etc make their way into somehow being interpreted as factual or predictive of a groups behaviors.

Just because 90% of X people in Y group do Z thing, doesn't mean you can assume that people of group Y ALL do Z thing and/or would continue to do Z thing or even be a part of that group for long. The 8% false positive or false negative ratio still equates to hundreds of thousands of people being misrepresented or made assumptions on.

This is a measurement of factors at a specific slice of time and shouldn't be used to predict the future. It is a present, measurement tool. Not a future predictive tool to be used to define inherent truths or anything else other than "people right now who look like this generally are in this political group". Did they do any checks to verify that this doesn't change over time? Did they cross check this with global results, or was this just a specific country, etc etc.

Don't let these articles fool you, stay curious, stay fair!