SoylentNews is people
SoylentNews is powered by your submissions, so send in your scoop. Only 15 submissions in the queue.
SoylentNews
A piece titled A.I. Astrology on Medium documents a new application available for IBM's Watson AI framework. The new application analyzes writing to build an estimation of the Big five personality index (which is a bit like Meyers-Briggs, but with more reliable results) of the writer as well as a few others. The writer of the article performed analysis on some famous figures.
You can play with the tool yourself here.
This discussion has been archived.
No new comments can be posted.
IBM's Watson for Personality Analysis
|
Log In/Create an Account
| Top
| 12 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
* m2 stares at the monitor... it looks like a hamburger...
m2 - that's a bad sign
(Score: 0) by Anonymous Coward on Friday August 14 2015, @12:42PM
Still just as meaningless in isolation as Meyers-Briggs and not necessarily better.
The problem is people have known about these indices for such a long time that they have integrated their personality to the point that many can game the results.
Let me write 5 different 6000 word essays designed to simulate a specific personality and with a bit of practice I doubt it would be a challenge.
Now let Watson take five different turing tests this way. Personality was originally part of the test and has been ignored for a long time.
(Score: 2) by ikanreed on Friday August 14 2015, @02:02PM
I'd like to hear your arguments here about why big five is "just as bad" as Meyers Briggs.
MB never really went through any sort of real scientific development. It started as conjecture, had a dash of research, then immediately crossed into being sold as a way of identifying things about employees. Later research then started to indicate it was unreliable, didn't predict much, and needed serious refinement.
That refinement eventually led to a richer personality spectrum, Big Five being a major output of that research.
Now, this particular technology is of questionable quality, because it's just using a neural net to link arbitrary inputs to arbitrary outputs, and isn't really different from me reading what you posted here and going "Yep, you're clearly pretty critical." It's just mechanized intuition.