Computer Faces Human Psychological Test

A reader writes: " The test known as the MMPI-2 (Minnesota Multiphasic Personality Inventory) is going to be administered to a computer-based personality. The test subject is GAC - Generic Artifical Consciousness." This is the first time I've heard of GAC - but this test is quite different from the Turing Test.

Comments

The acronym

[hugg]

G)enerator of A)rrogant C)laims.

seems a bit ./'ed... (mirror)

[gimpboy]

here's the article.


For the first time, a standard psychological test used by clinicians worldwide in the evaluation and treatment of adults will be administered to a machine-based artificial personality.

The test is known as the MMPI-2 (Minnesota Multiphasic Personality Inventory).Developed as a specialized psychological test for the measurement of psychopathology, the MMPI has been the preferred psychosocial diagnostic instrument among clinicians for the past 50 years.

Originally published in 1940 by Hathaway and McKinley, MMPI has been implemented in many clinical and non-clinical contexts, including medical, educational, medicolegal and organizational settings. A restandardisation and partial revision of the MMPI resulted in the publication of the MMPI-2 in 1989.

GAC (Generic Artificial Consciousness -- pronounced "Jack") is the artificial personality being developed at the Mindpixel Digital Mind Modeling Project with the collaboration of nearly 40,000 Internet users from more than 200 countries worldwide.

GAC will be evaluated using the MMPI-2 over the next several months to assess its learning of human consensus experience from the Mindpixel project's large and diverse group of users from many different cultures.

The test will be supervised and interpreted by Dr. Robert Epstein, one of the world's leading experts on human and machine behavior.

"Nothing like this has ever been attempted," said Epstein. "We're evaluating thousands of people worldwide as if they were one collective individual."

"We don't know if it is possible to build a normal personality out of millions of little pieces. This experiment will tell us how reasonable the idea is," Epstein added.

In the nearly one year the project has been online, Mindpixel's Internet contributors have made nearly 8 million individual measurements of more than 355,000 individual items of human consensus experience.

The project's organizers hope that they will gain enough information by the time the project's data collection phase is complete (2010) to build a highly accurate statistical model of an average human mind which they hope can be used as a foundation for true artificial consciousness.

One of the world's leading experts on human and machine behavior, Robert Epstein received his doctorate in psychology at Harvard University in 1981. He is Editor-in-Chief of Psychology Today magazine and University Research Professor at United States International University in San Diego.

He is also the founder and Director Emeritus of the Cambridge Center for Behavioral Studies in Massachusetts and Adjunct Professor of Psychology at San Diego State University. He was also the former director of the famed Loebner Prize competition in Artificial Intelligence.

The Mindpixel Digital Mind Modeling Project was launched on July 6, 2000. It is the world's largest Artificial Intelligence effort, with nearly 40,000 contributing members in more than 200 countries.

[Contact: Dr. Robert Epstein, Christopher McKinstry]

02-Jul-2001

use LaTeX? want an online reference manager that

Re:MMPI

[prizog]

rodentia [hey, wasn't she spammer?] writes:
"Finally judgements about normativity, while they strike some as distasteful, are really the only way to generalize about psychological phenomena."

Certainly not. Philip K. Dick said "Reality is that which, when you stop believing in it, doesn't go away." One measure of sanity is whether your conception of reality agrees with the actual reality. Certainly, in a society of delusional types, a realist should not be considered mad.

r:
"The idea of universal sanity or a generalized idea of madness are simply absurd and possibly dangerous."

Certainly, they are not as dangerous as a culturally influenced definition of same. Just ask a gay person.

MMPI

[hodeleri]

The MMPI is quite comprehensive. It is designed so that it is effectively impossible to hide anything. My psych professor said that he took one and (knowing how the test was designed) he tried to cover up that he had a major traumatic event in his life (which he did). The test showed it anyhow.

The test won't, however, say that you are a serial killer or homicidal maniac...

Re:MMPI

[rkent]

And for this reason, the MMPI is actually a ridiculous test to give to any AI system at this point. Have you ever READ the MMPI? I'd link to a copy online, but it's carefully guarded intellectual property, and costs hundreds of dollars to administer and score. But here's the NCS page with a Q&A about it if you want some particulars.

That said, it's fucked up. One of my roommates had to take it once (aside: we moved him out the next semester:), and he let us read it. There are 567 "yes/no" questions, and the first one is "I enjoy reading sports car magazines." Others include such obviously gender biased items as "I sometimes wear women's clothing." Basically, my point is, it's a very US-centric test designed to "catch" people who think differently than our definition of "normal." I suspect that the AI device wouldn't even have a coherent answer to many of the questions, since it's never lived in American culture, and doesn't have a childhood to be analysed as "traumatic," etc etc.

I can't see anything useful coming from this except perhaps a false reassurance that "this system won't go nuts like HAL-9000."

(yes, this is posted under another message below, but it fits better here.)

---

Re:MMPI

[rodentia]

I agree that giving this test to an AI system is totally marketroid hype. I'd like to clarify your view of the MMPI, having taken it twice.

Yes, some of the questions do seem a bit bizarre and it is often difficult to concieve either the purpose of the question or how to answer, but there is tremendous redundancy built into the test and it is scored on a curve. That is, answering yes to question 241 does not tag you a psychopath; ultimately the results are correlated against the sampling of individuals who have taken the test, as is the Rorshach. The tested individual is found to have tested within some standard deviation along nine (IIRC) scales. Individuals with particular patterns of results can be associated with others with similar patterns. Occasional questions are "hot buttons" designed to illicit responses which indicate that the subject has poor reality checking or exhibits dangerous behavior. Of course, the test is only useful in conjuction with clinical, therapeutic observation.

Finally judgements about normativity, while they strike some as distasteful, are really the only way to generalize about psychological phenomena. The range of normative behavior, while flexible, is the only standard available. The idea of universal sanity or a generalized idea of madness are simply absurd and possibly dangerous.

Re:And the purpose of this is what...?

[mindpixel]

IRTP: I run this project.

You are right, GAC is one heck of a psychological test on its own. However, unlike the MMPI, GAC has not explicitly been tested with both normal and clinical populations, so it would not be good at distinguishing the two.

The whole point of giving the MMPI to GAC now is to see what type of personality you get when you blend together 40,000 personalities. Looking at the data as it comes in (after recovering from the slashdot hit that killed the site for a while), GAC looks a little depressed and slightly sexually deviant. GAC also is showing some strong somatic symptoms (complaints about the body) which is of course quite ironic considering it doesn't have a body. But other than that, GAC appears normal. Quite human.

GAC Not exactly too clever (example)

[CyberKnet]

And the winner is:

1. Question asked of GAC 07/01/2001 16:03
2. Registered Login Required [href]
   --------
   I think the answer to: Donating $2.00 via PayPal is a good way to contribute to the mindpixel project is:

   FALSE

---

And the purpose of this is what...?

[kriemar]

The site is slashdotted, so it's difficult to determine what exactly is going on.

About the beauty of the MMPI: the MMPI is empirically validated, meaning that its scales were developed by selecting items that correlate with an objective criterion. Because items might correlate with an objective criterion (e.g., homicide) without looking like they would, you can get "subtle" scales--scales that are difficult to fake responses to.

But what's the point of giving GAC this?

Say the test predicts the computer is a psychopathic deviate. Does this mean the GAC is not functioning like a normal individual? Does it mean the GAC is psychopathic? Does it mean the average internet user who supplied observations is a psychopathic deviate?

What's ironic is that GAC may actually be a more comprehensive psychological resource than the MMPI. By selecting observations from so many individuals, GAC has tremendous data on what typical thought patterns are.

In fact, GAC could be used as a psychological test itself. This may be why they are giving GAC the MMPI. If GAC is "normal", so to speak, you could use GAC as a test. Even if it's not normal, you could "adjust" GAC to have responses more in line with a "normal" MMPI profile.

Then, you could "test" individuals by having them converse with GAC. The extent to which you "agree" with GAC or not is a gauge of the deviance of your personality. Ideally, it would be just like having a conversation with someone, only the extent to which you get along with GAC reflects on your abnormality.

Interesting...

Re:Hello. My name is Eliza. What's your problem?

[dasmegabyte]

"Dr. Sbaitso" _WAS_ Eliza. Same algortihm and everything...only named differently so as not to pay any royalties to the copy of "101 BASIC Computer Games" they lifted it from. They stuck in a stupid age check if you tried to talk to the doc about sex...which sucks, because I've a feeling that if my Thunderboard could have helped me through my youthful urges, I wouldn't be the cross dressing deviant I am today.

What is the Mindpixel Corpus?

[Neon Spiral Injector]

From Mindpixel's website:

The Mindpixel Corpus is the world's first and largest database of validated consensus human knowledge. It is an ever improving approximation of the mind of an average Internet user, constructed and owned tens of thousands of people just like you from all over the world, speaking all of the world's languages.


Emphesis added by me.

So, this thing is going to be obsessed with eBay and pr0n.

--

Democratic intelligence

[Shotgun]

"We don't know if it is possible to build a normal personality out of millions of little pieces. This experiment will tell us how reasonable the idea is," Epstein added.

I predict that the result will be about as useful as Congress.

advertizing

[Refrag]

You've gotta love this stealth advertizing for Spielberg's AI!


Refrag

Hello. My name is Eliza. What's your problem?

[hillct]

Didn't we try this 30 years ago? Did it work then?


--

Re:Hello. My name is Eliza. What's your problem?

[Anonymous Coward]

Why do you think it didn't work then? Tell me about your 30 years ago.

That broken link

[General_Corto]

Is fixed here.