Re: 3 "Real" Conscious Machines [WAS Re: Singularity: A rock 'em, shock'em ending soon?]

From: Ben Heaton (factitious@gmail.com)
Date: Wed Jan 18 2006 - 02:02:16 MST


On 1/17/06, Woody Long <ironanchorpress@earthlink.net> wrote:
> There IS an implied test in the Searle Chinese Room task. The Searle
> machine consciousness test is in the form -
>
> 1. The man in the room understands the incoming language like this i.e.,
> receives/processes it as a human level consciousness like this.
> 2. The system in the room has been evaluated to be receiving/processing it
> in this exact same way as this human level consciousness is.
> -------------
> 3. The system has passed the Searle test - it is exhibiting some machine
> consciousness.
>
> OR
>
> 1. The man in the room understands the incoming language like this i.e.,
> receives/processes it as a human level consciousness like this.
> 2. The system in the room has been evaluated to not be understanding
> (receiving/processing) it in the same way as human consciousness does, but
> to be just heuristically shuffling data cards without any understanding of
> what it is shuffling
> -------------
> 3. The system has failed the Searle test - it is not exhibiting machine
> consciousness.

So by evaluating whether the system understands its input, you can
determine whether or not it's conscious. The question is, how do you
go about evaluating that? This is the crucial part in constructing a
test, and it's not obvious to me that it's an easy matter to resolve.
Do you have any ideas on how to devise such a test?

-Ben Heaton



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT