The story reported it as an 18 month level so I would take their word for
it until proven otherwise. No need for the interaction to be typed. The
person interacting could use voice recognition programming, but that is a
trivial point. If we take a dialogue between, let us say, 10 human
adults with 10 real human 18 month-olds and 10 human adults with "Junior",
as we will call this R4P model, and type the 20 dialogues into an online
discussion, then we can do a Turing Test. The statisticians can evaluate
the results. That will give us the proof though I think we'll have to
spiff up the research design a little first.
POC: What Baby Junior wanna be when Junior grow up?
Junior: Wanna be president.
POC: Baby Junior sure?
Junior: Sure as steppin on a texas cowplop in June, hombre.
POC: Baby Junior have to go to school.
Junior: What be school?
The staff of Saturday Night Live will be the judges who will sort the
protocols into two piles, (1) rated as dialogue with R4P-Junior; (2) rated
as dialogue with a human 18 month old.
R4P-Junior would go to robot school where robopsychologists would program
it, developmental step by developmental step up to adult levels of verbal
On Tue, 21 Aug 2001, Zero Powers wrote:
> >From: Party of Citizens <firstname.lastname@example.org>
> >Have they passed the Turing Test at the 18 month level as reported in the
> >story or not?
> No. It would be pretty hard to even conduct that sort of test seeing as how
> there aren't many 18 month old humans who can even do IRC, much less give
> any sort of coherent response to a typed question.
> "I'm a seeker too. But my dreams aren't like yours. I can't help thinking
> that somewhere in the universe there has to be something better than man.
> Has to be." -- George Taylor _Planet of the Apes_ (1968)
> Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:40:12 MDT