[aiethics] Let's begin...

[Noel Bush]

Alicebot AI Ethics Committee - http://www.alicebot.org


Pedro wrote:

> I think a clear statement of policy is needed from the foundation
> before to start on the purpose and goals of this forum in order to
> refine it and elaborate on details.
>
> Otherwise the debates might lead to nowhere.

Hey Pedro, I agree with you w/r/t other committees, but I think that the
ethics committee is a real exception to this. I think that this is one =
where
the floor is wide open, and should remain so for some time. I think that =
the
AI ethics discussion will work *most* fruitfully as a sort of =
free-for-all,
where all manner of messages fly about, ranging from paranoiac fantasies
about bots taking over the world to carefully reasoned arguments, heated
personal arguments, and the whole nine yards. I think that a discussion =
of
ethics can't be reasonably constrained by anything! This is the place to
"let it all hang out". If the ethics mailing list were a room, I would =
put a
big "BEWARE" sign on the door. This ought to be the list that will flood
your mailbox with hundreds of messages a day!!! :-)

Really, I'm serious. This is the kind of topic that stirs up lots of =
energy
in people. Everyone has opinions about ethics. Every opinion has some
implicit ethics. Discussions of ethics are by nature irrational and
anarchic. I may be subverting the Chair by saying all this :-), but I =
would
encourage a completely "rhizomatic" (cf. Deleuze & Guattari or
www.rhizome.org if you prefer) discussion on this list. I think this is =
an
"open space".

"Open space": Rather than getting myself a "proper" education, I spent =
my
meager years of higher education in a zany and credentialless =
experimental
department of tiny Bard College in a program called "Music Program =
Zero".
The entire curriculum of this department depended entirely on
self-organization. A key component of the program was something called a
"commonspace", which was a weekly meeting open to the whole community
(meaning students and anyone) in which people would appear and, um, do
something. I say "um, do something" because in this space there was =
rarely a
fixed agenda. One time I arrived early and brought gigantic sheets of =
black
plastic which I used to turn the whole space into a lightless cave. =
Everyone
who arrived was greeted with a warning sign on the door, and when they
entered they spent 2 hours sitting in the dark with the complete =
recordings
of Blind Willie Johnson playing at maximum volume. There was no =
discussion
afterwards, no metatext to give everyone a set of instructions about how =
to
think or what to think about. Afterwards people were variously angry,
befuddled, inspired, blase, etc. The professor branded me a sadist. My =
aim
had simply been to create a very immersive listening experience. The =
social
aspects were frightening, sitting in the utter darkness with strangers
listening to this music that covers a wide range of intensities and
topicalities and banalities and profundities within a very narrow sonic
expressive bandwidth. I'm not sure what the outcome of this was -- I =
mean
that nothing in future discussions was directly "traceable" to =
particular
"inputs" from this event. But in its radical refusal to be explicit in =
its
discourse and total in its ontological prioritization of *experience* =
over
description it typified the work of that program.

There were for sure ethical issues being addressed in work like that. In
fact many "issues" were addressed -- also a fair amount of time wasted =
--
but the absence of any regulated process with an associated =
hard-and-fast
risk/reward ratio brought with it the possibility, occasionally =
realized, of
making more profound progress than any academic structure would permit.

On the other hand, as the person responsible for organizing the =
day-to-day
development work of a company that had about ten product development =
efforts
and numerous customer projects going on simultaneously, and a
frequently-shifting set of priorities for the same, I had to organize
development team work in a way that would produce real results, and =
fast,
and within budgetary constraints. Never mind that most of those results
never got out the door; nonetheless work had to get done and it did. =
What I
frequently noticed in those experiences was that many good ideas had to =
be
tossed out the window. Part of my job, in fact, was to try to make
on-the-spot evaluations of whether the idea someone was proposing was
relevant to the task at hand, and if not, to quickly "can it" before
resources allocated to the project were wasted. I found this a
not-so-pleasant task, because in the corporate setting there is usually =
no
outlet for creative energies.

Very little of interest was addressed, consequently, except a business =
plan.

Alice is a unique situation because everybody involved is doing it =
freely,
purely on the basis of self-motivation. Nobody is employed by Alice Inc. =
We
still have the same need to channel energies, though. The architecture
mailing list became moderated precisely because it rapidly deformed into =
a
collective stream of consciousness in which few ideas were traceable, =
very
little of an agenda clear. Even in its reincarnation it's still bursting
with ideas and will need to have a more explicit agenda set soon.

But the ethics list is, I think, the place for a lot of that "excess"
energy. Most of the discussions we can have about ethical issues are =
purely
speculative. We are not at this moment threatened by Matrix-like =
extinction
from superbots. It is still possible to slam the door when the Grim
Singularity comes calling. As Rich said in our published conversation, =
there
are far more serious and pressing issues in the world today than what to =
do
about bots. That doesn't mean this conversation isn't interesting; quite =
the
contrary, as evidenced. And when bot technology becomes more =
sophisticated
and more pervasive, we may indeed have something to worry about. In the
meantime, this is a great place to exercise our minds, our capacities =
for
imagining all kinds of weird possibilities, even to expose our fantasies
about what those of us interested in bots are really after. From all of =
this
self-organization and mutual titillation and provocation and so on ought =
to
come not hard and fast results, but attitudes -- attitudes with clearer
shapes, deeper convictions.

Noel