From: Woody Long (ironanchorpress@earthlink.net)
Date: Sun Oct 23 2005 - 14:57:28 MDT
> [Original Message]
> From: Michael Vassar <michaelvassar@hotmail.com>
>
> There should be ways of confidently predicting that a given machine does
not
> have any transhuman capabilities other than a small set of specified ones
> which are not sufficient for transhuman persuasion or transhuman
> engineering. It should also be possible to ensure a human-like enough
goal
> system that you can understand its motivation prior to recursive
> self-improvement.
As a business owner whose business is to build / have built an android
robot and enter it in the Roboprize Prize Fight against a human-controlled
robot, I am faced by these SAI safety issues daily. The above sums up
precisely the company policy on android safety protocols. Androids of
course have humanoid or human-like strong AI embedded systems, and so are a
part of the SAI field.
It appears androids in households will first appear in Japan. Its predicted
that in 10 years, 1 of 3 Japanese households will have some sort of human
intelligent like humanoid in the house. Maybe in 20 years they could be SAI
androids. Thing is they can be connected to the internet and so are quite
powerful.
So my question is - in this context - what does a friendly android, or in
Japanese principle of harmony terms, a "safe'built" android mean,. i.e.,
what are the industry android safety protocols?
For me, there are three things the Japanese corporate builders don't want,
for fear of being sued and disliked by consumers. This is a huge cash cow
and all they want is to be "admired and desired" by consumers. So there it
is, in the consumer's house, stockholders awaiting the explosion of
consumer love, millions invested in development and brand naming, and for
sake of argument, let's say it is interfaced to your system. Now there's
motivation to make sure its not a defective product. These three primary
things are destruction of property, persons, or pets. So the android safety
protocols must be: androids can't - by goal, end or means - break the
stereo, harm the kids, or harm the pets. To be thus friendly and
safe-built, to have a perfect industry safety record will ensure consumer
love and desire, and profits.
This might not be the whole of the android safety protocols, but it is a
core, necessary part, for avoiding lawsuits and consumer contempt. The
android can't be sued for failing to save my life, it no such product
feature is claimed.
So this is why I agree with your following -
>An AI which others can see the non-transparent thoughts of, and which
others can manipulate the preferences and desires of, is in some respects
safer than a human, >so long as those who control it are safe.
In conclusion, this is why I feel confident in the prospect of friendly SAI
androids, and the growth of this industry.
Ken Woody Long
CEO
www.artificial-lifeforms-lab.blogspot.com
This archive was generated by hypermail 2.1.5 : Tue Feb 21 2006 - 04:23:16 MST