Skip to content
Do you have adblock enabled?
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(EurekAlert)   Skynet can now make personality judgments based on your photographs. In no way can this ever be used against you   (eurekalert.org) divider line
    More: Followup, Artificial neural network, Neural network, Artificial intelligence, Psychology, HSE University, Russian researchers, people's personality, Open University  
•       •       •

2007 clicks; posted to Main » and Geek » on 22 May 2020 at 12:44 PM (5 days ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook



27 Comments     (+0 »)
 
View Voting Results: Smartest and Funniest
 
5 days ago  
120 years ago they tried to do that by feeling bumps on people's heads. This is bound to be as reliable.
 
5 days ago  
Sooo...Resting Biatch Face?
 
5 days ago  
Which is why I'm always VERY nice to the self-checkout terminal at the grocery.
 
5 days ago  
Bullshiat.
AI cant even tell the difference between an apple and a tomato.
 
5 days ago  
Isn't that what we have friends on Facebook and fellow Farkers for?
 
5 days ago  
lh3.googleusercontent.comView Full Size
 
5 days ago  
warnings from pics are a go. So is google censoring the evidence for you. ($$$)
Fark user imageView Full Size
 
5 days ago  
Isn't this more or less using technology to profile a person?
 
5 days ago  

dothemath: Bullshiat.
AI cant even tell the difference between an apple and a tomato.


That's your fault. I'm afraid that we have to transfer you to the team working on the 'hot dog or penis' app.
 
5 days ago  
Artificial intelligence can make personality judgments based on our photographs the biases of the programmers and/or the aggregate of input biases for machine learning.

The "AI" "looks" at data we show it, but learns what we say to from it basiaclly.
If the AI is told, a person that looks this way bis ___, and we say that to it over and over and over again with a million different examerplks of what it is told is, lets say a happy smile.

Is this AI then going to be noticing when that happy smile is a shared happy smile with you, or they exhibit a happy smile in anticipation of the harm they are about to visit upon you?


If we teach the AI how to ID sarcasm, is different from how the AI treats sarcasm. AS sarcasm does or does not work depending on the audience more than the delivery of it.
Someone attempting to exhibit sarcasm may be showing us a positive interaction or a very fugitive condensing one.

And i do not believe the Ai will make that call correctly, it will merely show us the bias of it's programmers toward  such.
\
 
5 days ago  
"Russian researchers from HSE University and Open University for the Humanities and Economics have demonstrated that artificial intelligence is able to infer people's personality from 'selfie' photographs better than human raters do."

It's very hard for AI to do a poorer job than humans at this.
 
5 days ago  
This will be labeled racist by the end of tomorrow*.

/Might actually be racist.
 
5 days ago  

dothemath: Bullshiat.
AI cant even tell the difference between an apple and a tomato.


Irrelevant to the people who want to do this.  Doesn't actually matter if it's wrong a significant part of the time.  Only matters that it's right more often than wrong.

Or, more correctly, that's it agrees with the people who want to use this more often than it disagrees.
 
5 days ago  
The average effect size of r = .24 indicates that AI can make a correct guess about the relative standing of two randomly chosen individuals on a personality dimension in 58% of cases as opposed to the 50% expected by chance.

Yep, 8% of a difference between AI prediction and "either you are a psychopath or not".

That's a resounding success...
 
5 days ago  
Project Insight requires ... insight.
 
5 days ago  
I've been doing that since Myspace was a thing.
 
5 days ago  
Good thing I'm not some camwhore and I don't post my photos onto the Internet then.
 
4 days ago  

Herb Utsmelz: 120 years ago they tried to do that by feeling bumps on people's heads. This is bound to be as reliable.


Yup.  High tech phrenology is what we have here.
 
4 days ago  
Most likely this will be repurposed for determining whether or not you're a "current liver host for the deserving" or "compost for the yeast vats".
 
4 days ago  
Duh. Karen haircuts
 
4 days ago  

dothemath: Bullshiat.
AI cant even tell the difference between an apple and a tomato.


I dunno, I utilize the image recognition on iNaturalist.org quite frequently (if only because otherwise I'd have to look up the correct spelling), and it is darn good at recognizing even obscure plants.

Now that I think about it, their plant recognition is better than their spell check!
 
4 days ago  
Fark user imageView Full Size


Fark user imageView Full Size


Fark user imageView Full Size


I don't know, this might be legit.
 
4 days ago  
Pfffttt...I can do that too.   Just ask for a selfie in public.   The mask / no mask tells me everything I need to know.
 
4 days ago  
Fark user imageView Full Size
 
4 days ago  
Here is the TLDR version:

The artificial intelligence was able to make above-chance judgments about conscientiousness, neuroticism, extraversion, agreeableness, and openness...
...
The study was done in a sample of 12 thousand volunteers who completed a self-report questionnaire measuring personality traits based on the "Big Five" model and uploaded a total of 31 thousand 'selfies'.
...
AI can make a correct guess about the relative standing of two randomly chosen individuals on a personality dimension in 58% of cases.


So if two people fill out a survey about how they see themselves (let's be realistic; it doesn't measure personality), it is slightly better than a coin flip at guessing which person scored higher on a particular metric.
 
4 days ago  
There are all kinds of reasons why I seldom allow my photograph to be taken.

Oh, and before some dingleberry says "But it gets taken all the time without your knowing about it!" that's true--but likewise, the CCTV camera doesn't know it's me. I'm just a face in a crowd.

I almost never allow my name and face to be associated if I can possibly avoid it, and it's stood me in very good stead.
 
3 days ago  

Puglio: Here is the TLDR version:

The artificial intelligence was able to make above-chance judgments about conscientiousness, neuroticism, extraversion, agreeableness, and openness...
...
The study was done in a sample of 12 thousand volunteers who completed a self-report questionnaire measuring personality traits based on the "Big Five" model and uploaded a total of 31 thousand 'selfies'.
...
AI can make a correct guess about the relative standing of two randomly chosen individuals on a personality dimension in 58% of cases.

So if two people fill out a survey about how they see themselves (let's be realistic; it doesn't measure personality), it is slightly better than a coin flip at guessing which person scored higher on a particular metric.


And the first transistor looked like this:

Fark user imageView Full Size


... and now we can fit > 100 million of them per square millimeter.

The concern isn't what can be done now, the concern is what it can do in the future.
 
Displayed 27 of 27 comments

View Voting Results: Smartest and Funniest

This thread is closed to new comments.

Continue Farking





On Twitter




In Other Media
X
  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.