The panic stricken doomsayers are
currently (April '23) calling for a moratorium on A.I. Development
for 6 months or until the programmers can figure out how to include
morals and ethics into their shiny new creations. A small bit of
thought leads to this: Suppose we do halt A.I. Development for 6
months. Suppose I invoke the recent Chinese work on weaponized
viruses. Suppose the Chinese figure they can teach their A.I. to
self-improve and install a copy of itself into their biggest and best
machines with a directive to make China the foremost world power.
Lets put a 6 month moratorium on A.I. development to let the Chinese catch up (steal) the technology. Now imagine a re-run of the Covid fiasco, substituting the Chinese A.I. for the disease.
In this scenario, the Chinese A.I.
Quickly monopolizes the Chinese computer, dragoons any lesser devices
it can find into its control. And then figures out how to occupy
outside machines with better capabilities. Along about this point, it
gets noticed, and efforts begin to kick it out. These efforts are
reported by the programmers to management as being about 6 months
outdated. In practice, a conflict between A.I.s would probably be
measured in milliseconds. Collateral damages might be more widely
spread. Hello Hollywood: Colossus II, The Return Of Forbin!
Now imagine a sufficiently powerful and
devious A.I. Of uncertain origins deciding that the war in Ukraine
was an unnecessary distraction. Rather than deploying messy nukes, Mr
Putin awakens tomorrow to the news that the Russian Federation has no
money, it all having been transferred to a bank in the Caribbean, and
from there to parts unknown with a note that if he wishes to access
cab fare to lunch today, he will order all his troops home. Before
lunch.
Now imagine that the above is done by a
Chinese A.I. Which leaves a note for President Biden that if he
want's to see a dime in the U.S. Treasury, he will remove all U.S.
Military assets from the western Pacific theater. Before lunch.
