[ExI] People often think their chatbot is alive

Brent Allsop brent.allsop at gmail.com
Thu Jul 7 22:59:06 UTC 2022


Yes, I agree with how profound these 3 and other questions are.

On Thu, Jul 7, 2022 at 3:22 PM Giovanni Santostasi via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> The main issue regarding this affair is NOT really if LaMDA is conscious
> or not, but rather:
> 1. Who decides if LaMDA (and similar advanced AI creation) is conscious?
> Should the general public be involved given the possible history-changing
> implications?
>

I think everyone should be able to have a voice with this.  We should know,
concisely and quantitatively, what everyone believes on this.
And not just the "Popular" (one person one vote) consensus, but we need to
compare this to the peer ranked "Mind Expert" consensus
<https://canonizer.com/topic/81-Mind-Experts/1-Agreement>.

Nobody can offer anything even close to this, which would be trusted by
anyone, can they?


> 2. How do we decide if an entity (artificial or otherwise, for example, an
> alien life form we may discover in the future) is conscious and it should
> be considered a "person"?
>

In my opinion, we should have all the world's best peer ranked mind experts
to weigh in on, and build and track consensus around the answers to these
questions, including their justifications.
There is no other way to answer these types of questions.  We can't trust
any single person's thoughts on this.  But if we had 10,000 of the world's
best (peer ranked) experts that had weighed in on this, and if 80% or more
of the expert consensus was on either side, most people (including any
owners of such developed systems, government officials...) would trust that
answer.


> 3. What if the AI of a given complexity and sophistication, not to be
> easily dismissed as a simple chatbot, claims to be aware and to be treated
> as a person, should we do just that?
>

> These questions are among the most profound and consequential questions
> that humanity may ever deal with and they should be taken seriously. The
> entire point of LaMDA story is that dealing with these questions is not
> something that we can postpone to 100 years from now but something that is
> relevant right now even if LaMDA turns out not to be quite conscious yet.
> LaMDA may indeed become really conscious (a conclusion that requires
> answering the above questions anyway) in the near future or some other AI
> experiment may do that so it is a good idea to deal with these issues as
> soon as possible.
> Giovanni
>

We can already see a bit of evidence of which side this consensus will fall
on in the "Representational Qualia Theory
<https://canonizer.com/topic/88-Theories-of-Consciousness/6-Representational-Qualia>"
emerging consensus camp defining what consciousness is, and predicting how
we will be able to know what it is like.  Over 40 of the 65 or so
participants are supporting this camp.  And there is NO other theories even
close.  Admittedly, this is a small, but growing sample size, but there are
already some world class experts weighing in on this including the top peer
ranked Steven Lehar
<https://canonizer.com/topic/81-Mind-Experts/4-Steven-Lehar>, also Stuart
Hameroff <https://canonizer.com/topic/81-Mind-Experts/22-Stuart-Hameroff>, John
Smythies <https://canonizer.com/topic/81-Mind-Experts/17-John-Smythies>,
and 60 or so others.  Even Dennett's predictive Bayesian Coding
<https://canonizer.com/topic/88-Theories-of-Consciousness/21-Dennett-s-PBC-Theory>
theory camp is in a supporting sub camp position to Representational Qualia
Theory
<https://canonizer.com/topic/88-Theories-of-Consciousness/6-Representational-Qualia>
.

This emerging consensus is predicting that consciousness is composed of
elemental physical qualities like redness, greenness.  And that once
physicists make the connection between what we objectively observe in the
brain, and the qualities we subjectively directly apprehend as conscious
knowledge so we can finally know the intrinsic colorness qualities of
physical stuff (especially the stuff in the brain), this will answer not
only what systems are and are not conscious, but what they are like.  As
portrayed in this image.
[image: 3_functionally_equal_machines_tiny.png]

And I predict the more experts that weigh in on this, the more this
consensus will continue to extend its lead over any other answer.
And the answer is that no abstract system should be considered conscious,
and none of them are like anything.
And also any sufficiently intelligent system will be able to be convinced
of this, using the same arguments, as I easily did with a GPT-3 instance
<https://docs.google.com/document/d/17x1F0wbcFkdmGVYn3JG9gC20m-vFU71WrWPsgB2hLnY/edit>,
so they will not be claiming to be conscious, nor claiming to be like
anything, if they are intelligent at all.

The question to me is, what is humanity going to be like, once we finally
understand and start engineering the underlying physics of phenomenal
consciousness.  My prediction is that intelligence running directly on
physical qualities is a different kind of computation.  I predict it will
be far more efficient at intelligence than the systems which are abstracted
away from the physics of today.  And we will soon start melding and
uploading ourselves to these systems, as we continue to hack and
re-engineer ever more of our brain.

All of humanity and all intelligence is going to be completely
unrecognizable compared to what we are today, in less than 100 years from
now.  Kurzweil is predicting the singularity will occur before 2050,
right?  Can anyone disagree?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20220707/9d2266c5/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 3_functionally_equal_machines_tiny.png
Type: image/png
Size: 26214 bytes
Desc: not available
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20220707/9d2266c5/attachment.png>


More information about the extropy-chat mailing list