[ExI] Betting on the end of the world

[Adrian Tymes]

On Tue, Jan 17, 2017 at 1:08 PM, Dan TheBookMan <danust2012 at gmail.com> wrote:
> http://econlog.econlib.org/archives/2017/01/my_end-of-the-w.html

If that's what it takes to save the world, then I for one consider the
world saved.

And with that addressed, can we go on to addressing more likely ways
humanity can use AI to mess itself up unintentionally?  For example,
this unemployment issue - is it possible to develop robo-missionaries
who can go among the great displaced populations, to the many
neighborhoods and small towns that are numerous in contrast to our
small number (setting aside the matter of identifying the small
fraction it would be unsafe for us to enter)?  Maybe preach to them
the ways and value of technical and vocational education, and
enlighten them into the wealth-generating portion of the labor force?
And while doing so, reveal to them how to spot and see through the
many deceptions their would-be masters have plied them with, and
beseech them to abandon their hate of other peoples who are not in
fact to blame for their misery?

I do not just mean some next generation chatbot, though that might be
a start.  Perhaps a step up might be a limited version of a Turing
Test: based on the very short chain of posts typical of a comments
thread, have the bot be indistinguishable from (at worst) a human
troll.  (Any organization capable of developing this is very unlikely
to be using it to promote the echo chambers currently seen in the
wild, therefore we can conclude this is probably not happening today.)
 Bonus points if, in most posts, it links to evidence for its
counterpoint.

But to truly have a massive effect, engagement with people offline is
likely to be required.  That requires a chassis to go and talk to
them, and one which they are willing to talk to.  A video screen on
wheels is not a serious effort at this, no matter what its backers
say.