Emotitoy / Emotion Gauging Software, We can't ban them from real life! Options Track this topic Email this topic Print this topic Download this topic Subscribe to this forum Display Modes Switch to: Outline Standard Switch to: Linear+

[Draco18s]

Here's the thing I heard about:
http://groupmedia.media.mit.edu/jk.php

And here's the story I heard it from:
http://www.sciencefriday.com/program/archives/201010151

They've got software, that without listening to the words of a job interview, could predict to within $1000/yr. the offered salary.

[Karoline]

Here's the thing I heard about:
http://groupmedia.media.mit.edu/jk.php

And here's the story I heard it from:
http://www.sciencefriday.com/program/archives/201010151

They've got software, that without listening to the words of a job interview, could predict to within $1000/yr. the offered salary.

Yes, but by how much did the salary vary? I mean, if you're talking about 10 applicants for Burger King, that's not difficult. If you're talking about 10 applicants for nearly any particular job, that's also not very difficult, because a particular job doesn't tend to have massive pay differences. Unless you're talking about a job that had a variance of 30k or more between offers, being with 1k isn't that impressive. I bet I could get within 2k at most interviews without even attending the interview based purely on the job.

And you can't say that it didn't know what the job is, because otherwise an enthusiastic Burger King manager is going to generate $200k results, and a sullen hospital head will generate $30k results.

And of course it is exceedingly difficult to tell what kind of a bonus it would give. Does having that thing with no real social skill make you as convincing as the world's most convincing person? Unlikely.

[Angelone]

(IMG:style_emoticons/default/rotfl.gif) Greatness I really just want to shout "Eat it haters!" but I won't because no one else is around.

[Draco18s]

Yes, but by how much did the salary vary? I mean, if you're talking about 10 applicants for Burger King, that's not difficult. If you're talking about 10 applicants for nearly any particular job, that's also not very difficult, because a particular job doesn't tend to have massive pay differences. Unless you're talking about a job that had a variance of 30k or more between offers, being with 1k isn't that impressive. I bet I could get within 2k at most interviews without even attending the interview based purely on the job.

Unfortunately they didn't get that detailed about it. But you're right.
They were however basing the information solely on body language. Just what they could pick up with an iphone, to be precise.

The other little toy (the Jerk-o-Meter) tells you how much of a jerk to (eg. how little you were paying attention to a conversation with) your significant other. When you should have been more excited, etc.

[Smokeskin]

Well, that the software can tell the outcome of the negotiation by observing it with no sound is hardly the same as giving a negotiator an advantage, especially not one as big as the emotitoy does mechanically in SR4...

[Draco18s]

Well, that the software can tell the outcome of the negotiation by observing it with no sound is hardly the same as giving a negotiator an advantage, especially not one as big as the emotitoy does mechanically in SR4...

It's "emotisoft in progress." It's the fact that the research is being done to get to "emotitoy" type software.

[Irion]

Well, the jerk o meter is a funny idea, but it ain't helpful in negotiations or something. Because everyone is able to present boring facts in an entusiastic voice. (Or close to everyone) And of course most people have the ability to sound interesteted, even if they are not.

[Draco18s]

Well, the jerk o meter is a funny idea, but it ain't helpful in negotiations or something.

Not yet anyway.

[Thanee]

"Eat it haters!"

If it is gone then, that seems like a valid course of action. (IMG:style_emoticons/default/biggrin.gif)

Bye
Thanee

[hobgoblin]

i suspect that with enough number crunching, one can even detect when someone is faking one response to mask another.

[Draco18s]

i suspect that with enough number crunching, one can even detect when someone is faking one response to mask another.

Given that two people who have never met lie to each other no less than 3 times (and as many as 12) in 10 minutes, probably.
(Various sources)

I think there's a show on Radio Lab, but I can't find the one I think I'm thinking of, where they video taped people introducing themselves and then having those people point out (privately to the researchers) each time they said something that was untrue and how untrue it was. IIRC that study found people lie as much as 30 times in 10 minutes (from "your hair is nice" to "I'm a rock star").

[hobgoblin]

Two things came to mind now.

1. people with arperger have a issue with eye contact, as such they are often unconsciously seen as habitual liars by people not in the know.

2. There will be autonomic clues. If not, the person would either be a sociopath or a nexus series android.