Re: UPL: Cautious Plans?

Keith Elis (hagbard@ix.netcom.com)
Sun, 23 Nov 1997 13:06:49 -0500


Twink wrote:

> At 12:00 AM 11/23/97 -0800, Michael M. Butler <mbutler@comp*lib.org> wrote:
> >OK, I'll formally mention (having elected myself as Devil's Advocate
> >here) that your plan does not include:
> >
> >A.) Handling publicity (or secrecy)
> >B.) Handling physical security: extra-lab *and* intra
> >C.) An explicit guarantee of pulling the plug if needed
> >D.) A description of what would be grounds for pulling the plug
> >E.) A description of what mechanisms would be used to pull the plug.
>
> All great ideas. I will have to include them.
>
> >This is all stuff that's standard for dealing with those microorganisms
> >you mentioned in a post to me.
> >
> >If you don't consider uplift at least as risky as working with
> >bioweapons... _why_?
>
> Because we are dealing with a multicellular animal. Much easier to
> control than current microbes. Also, the octopus is a marine
> animal. Despite its escape artist skills, it will be much easier to
> control such an organism, which is less likely to, say, escape to
> the oceans from my apartment than, say, an uplifted rabbit or dog.
>

With uplifiting, we face some of the same problems we face in dealing with
AI's. One of these similarities is our inability to actually know *WHEN* the
uplift has reached the level of sentience. (I.e., when do we *STOP* the
uplift?) Given that the uplift is likely to be a gradual process, we will
certainly have to find a method of estimating intelligence based on something
other than observed behavior or observed problem-solving ability. I suppose a
marked increase in brain activity/waves may be a clue, but this seems to be a
relatively inelegant, and in the end, inconclusive means of estimating our
success or failure. One concern that would stem from this is that the
newly-created intelligence may not *want* us to know it is intelligent. In such
a case, it is possible that we would have discovered the means to uplift
ourselves, but instead of doing so, we continue iterating our uplift procedures
on the test subjects, and before we realize it, we have uplifted them to be
smarter than we are.

Maybe this is not probable, but I think finding some way to answer the question
of when we stop the uplift is necessary before we can even begin.

Regards,

Keith