Using the latest technology in artificial intelligence and
super-computing, create software to make the worst
possible decisions.
These decisions can then be used to gauge how dumb
some
ideas are, in order to prevent actual people making bad
decisions. The proposals could be input into
the
computer
then assessed on the stupidity quotient of any particular
idea.
The computer, after crunching numbers and various
related
inputs, could then determine the SQ (Stupidity Quotient)
of
the idea - naturally the higher the SQ, the dumber the
idea
is.
For example, one could enter 'Build a wall between
Mexico
and USA and make Mexico pay for it'. In this case, using
the
huge database of information the computer would assess
the possible outcomes. After a few minutes crunching
some
numbers and extrapolating the results, this idea would
receive an SQ of about 140, or genius level stupidity.
In the case of the input 'Build a wall between Mexico and
USA' the AU might deliver a score of 120, or decently
stupid level.
In the case of 'Build a wall around Trump Tower' the
machine might explode or give a negative result. -20 or
maybe an OK idea.
The mechanism would include looking at historical data,
using the outcomes of similar decisions, to assess the
various likelihoods of different scenarios. The quantifier,
for the ultimate calculation of the SQ, would use many
different simulations, substituting the values of the
status quo, into the determinate factors of the historical
data.
Upon each iteration the score would accumulate,
resulting in the total SQ value. In the case a certain
outcome was bad historically, a positive value would be
accrued, and similarly for a good historical outcome, a
negative value could be accrued.
A simple algorithm would not suffice, and like artificial
intelligence the algorithm could self modify and adjust
according to various situations.
For example, if one were to ask AU if it is a good idea to
heat water in the microwave using a metal cup, the
program would look up the historical references on this
scenario, and also investigate some of the science behind
why doing such a thing, may or may not be a good idea.
Using reasoning and the ability to adapt, if the historical
data does not exist, the computer would attempt to
simulate the situation using science and mathematics. If
the goal (to heat water) cannot be achieved without
causing significant harm to the tool (microwave/cup)
then the idea can be assessed a positive score,
however if in the case of no harm being caused and the
goal being fulfilled, then a
negative score can be accounted for.
For a simple task such as heating a cup of water in a
microwave, the computer would give a resounding
positive score, may be also about 140-160 SQ, indicating
that it is indeed a bad idea.
For the way more complex situation of border control, a
great number of inputs could be assessed in a similar
manner, after each input has been balanced and checked
against historical data and artificially learned data - the
score can be accumulated and assessed accordingly.