• Welcome to the Internet Infidels Discussion Board.

Do we need to rethink the turing test?

 
 
 
L

Loren Pechtel

Super Moderator
Staff member
Joined
Sep 16, 2000
Messages
51,541
Location
Nevada
Gender
Yes
Basic Beliefs
Atheist
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
 
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

Well, really, what we need is a "chinese room test", in addition to a turing test, because I firmly believe both models are used by various people to varying extent.

There are the Chinese roomers, henceforth "Roomers", who hear something, say something, get a positive (fulfillment) response, and then reuse that. Or see something said by someone else, see a fulfillment response, and then reuse that. But no attention is paid to the mechanism by which the response fulfillment happens. Merely that "this is the response to this utterance" WRT subject and predicate.

Conversely, some people instead focus on operating a mechanism rather than invoking the magic of call/response.

I have my ideas as to which test someone of any given political persuasion would pass.

The thing is, Q creates an echo chamber wherein the use of various calls (democrats eat babies) are conditioned to positive response. It is the Roomers talking to other Roomers.
 
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

Well, an AI that can convincingly portray a Patriot Party flat-earther Karen may technically pass the Turing test, but is it really Artificial Intelligence?

On the other hand, and in all seriousness, i am getting much better at detecting very sophisticated robocalls. So, it's not that we need to rethink the concept. It's more like an arms race. The better the software gets at mimicry, the better the tester has to get to detect mimicry.
 
Worldtraveller said:
The irony being, I'm not sure LP could pass a Turing test. :D
Click to expand...

Honestly? LP is better than most. They have their issues (like kneejerk supporting violent cops and denying racism), but I have every hope that at some point, someone will figure out the thing that will finally help things click for him... I was really hoping the "why do you assume ageism and not viewpoint discrimination?" Would get him there, in the "50 year old man gets denied job" thread. Maybe it still will. Time will tell.

That said, he would absolutely do better to propose his solutions to such problems up front, though most people would. If he focused on procedural solutions rather than nitpicking about the exact source of the problem (especially when the solution would work for the symptom regardless of source), then he would probably be able to drastically improve the way people tend to view him.

Overall, though, don't see him failing a turing test (he is not non-intelligent) OR the Chinese room test (I don't think he is a roomer, either). ¯\_(ツ)_/¯
 
Keith&Co. said:
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

Well, an AI that can convincingly portray a Patriot Party flat-earther Karen may technically pass the Turing test, but is it really Artificial Intelligence?
Click to expand...

At least someone gets it!
 
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

How else would you formulate the test? I think it's a good test. The basic problem is philosophical. We have chosen to use human intelligence as a baseline for what intelligence is. Since humans are all that smart, it has problems. But if you open it up to other definitions of what it means to be intelligent, we quickly run into trouble. It's an infinite variety of what counts, and everybody wants to prove how their thing is the smartest, so use a definition where they shine. At least the Turing test does away with all that.
 
DrZoidberg said:
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

How else would you formulate the test? I think it's a good test. The basic problem is philosophical. We have chosen to use human intelligence as a baseline for what intelligence is. Since humans are all that smart, it has problems. But if you open it up to other definitions of what it means to be intelligent, we quickly run into trouble. It's an infinite variety of what counts, and everybody wants to prove how their thing is the smartest, so use a definition where they shine. At least the Turing test does away with all that.
Click to expand...

The thing is we have so many extremists these days that could probably be emulated with a few thousand canned responses and a weighted keyword system to decide which of them to pick.
 
Loren Pechtel said:
DrZoidberg said:
Loren Pechtel said:
All the total nonsense that people have accepted makes the bar for recognizing intelligence awfully low!
Click to expand...

How else would you formulate the test? I think it's a good test. The basic problem is philosophical. We have chosen to use human intelligence as a baseline for what intelligence is. Since humans are all that smart, it has problems. But if you open it up to other definitions of what it means to be intelligent, we quickly run into trouble. It's an infinite variety of what counts, and everybody wants to prove how their thing is the smartest, so use a definition where they shine. At least the Turing test does away with all that.
Click to expand...

The thing is we have so many extremists these days that could probably be emulated with a few thousand canned responses and a weighted keyword system to decide which of them to pick.
Click to expand...
But does the Turing test really count if the software design intent is to mimic someone who is using emotions, not reason? Just regurgitating canned talking points and favored accusations to live out their persecuted rebel fantasy?

This is more like a corollary to Poe's law than a failing of the Turing test. Actually, this already is covered by Poe's law.
without a clear indicator of the author's intent, it is impossible to create a parody of extreme views so obviously exaggerated that it cannot be mistaken by some readers for a sincere expression of the views being parodied
Click to expand...
i would suggest that software written to emulate an extremist would be a parody of that form of 'streemism.
FLATERTHBOT 9000 has no actual opinions, and is not learning software, just code to parrot the sounds the subject makes.
 
You must log in or register to reply here.
Back
Top Bottom