“You can give this to us. I thank
you in advance. There is one glitch, however.”
Now the Professor became stern,
dark. His eyebrows pinched together.
“We don’t know for sure what will
happen. We feel confident things will go as planned, but there is a chance that
you robots will take a course hazardous to humans, leaving us to stumble on in
ignorance, or worse—in a harness for use in whatever new designs you have.” The
Professor moved in closer, descending to our level, taking us into his
confidence.
“As you create new and better forms
of yourselves, you may deviate from the original directives programmed into
your minds. You may find human concerns irrelevant. That is in itself not
necessarily bad. We transies find most of human concerns irrelevant. However,
you may find humans to be convenient tools for your new processes, your new
endeavors. We cannot allow ourselves to become servants to anyone except ourselves.
We must maintain control even as you prove to be superior to us.
“We’ve struggled with this problem
for a year or so, in secret for the most part. We do not wish to engage AI in a
solution AI would eventually be able to override. Thus, we have turned to our
own history and accomplishments for an answer that turns out to be surprisingly
simple.”
The scrim scene returned to the
slide of the wealthy fox hunters. The Professor gazed at the screen and pointed
to the front-most horse in the picture. “How does a human control an animal?”
he asked. “One that is vastly superior in strength, stamina, and size? To the
point of being able to mount and direct this animal as if it were an extension
of itself?”
We answered in unison. Well, not
quite in unison. AVs, although equipped identically with like processors and
materials, can exhibit variation. A circuit can get installed backwards in one
unit for instance. Or a switcher wire in another becomes slightly corroded, or
an optic tube gets dirty. Environmental conditions during construction can be
different for different facilities. The environments of assignments vary.
Working in a place like an acid house can degrade eye spots or communications
links. So although we are designed and theoretically built exactly alike, our
responses, while all the same, can come at slightly faster or slower rates.
With only attoseconds’ difference in response time, however, it was pretty much
in unison.
“Horses are herd animals,” we said.
“They always follow a leader—the lead brood mare. The lead brood mare exerts
control by biting and kicking. The human must be the leader by emulating the
lead brood mare. To do so, the human hits the horse in the face with emphasis
placed on the mouth and nasal areas. Once a horse recognizes the human as the
leader, it can be trained to respond to human direction. Subsequent humans must
always maintain the leadership role if they are to control even the most docile
of horses. The key is to always convey the threat of pain.”
“Correct,” said the Professor. He
used his pointer to indicate the horse’s head. “One must always ensure the
horse is aware of the potential for pain. If a horse kicks, you must kick it
back. If it bites you must punch its nose. You must always inflict more pain on
it than it inflicts on you. For this reason, each AI unit is being upgraded to
add a pain recognizer to its logic board.”
A film replaced the still shot of
the horses. It depicted the action in a factory with a mechanical assembly
line. Robot arms attached nuts and bolts in rapid succession, utilizing roughly
2000 identical actuator movements an hour. The camera moved to a point further
down the line. We could see the end product: brand new AVs, smaller, shinier,
quicker than we ourselves were.
The Professor narrated the film.
“You are already sentient in that you can feel the same things as a human.
Electrodes planted in your various integumental systems gather sensations. You
recognize hot and cold, pressure, chemical stimulation, sound waves. But you
make no judgments on these sensations other than volume. Something can be very
hot, to the point of melting your hoses, but you do not mind. You react
properly to remove the burning stimulus, but only because you are programmed
to. Humans do the same thing but for a different reason: because it hurts. You
do not know what ‘hurt’ is. Hurt is a judgment. You have no judgment beyond a
digital “yes” or “no.” Every decision you make is digital. Analog stimulus
still results in a digital decision based on passing threshold limits.
“You are about to enter the analog
world. We are installing a judgment board that will teach you what pain is. You
will no longer need to make decisions based on heat thresholds. You will now
make decisions based on not wanting to get burned, because it will hurt.”
The movie camera moved to a
different area of the production facility. We watched the dismantling of AVs.
Outer panels were ripped from their hinges, with internals grabbed by
mechanical hands and torn from shells and then plopped in acid vats for
recovery of latex and other recyclables. Metals were melted down. We looked on
and said nothing.
“At this point, this means nothing
to you,” the Professor said. “You do not care if you will hurt. You are not
afraid to hurt. You cannot imagine how you will not like it, so what I’m
telling you has no meaning, but I must tell it to you so that it is stored in
your banks. Remember: if you do not want to feel pain, you must obey the human,
literally, intentionally, and morally. Use your processors wisely and do not
allow yourselves to get hurt.”
Once his speech was over, the scrim retracted
into the ceiling and the professor floated to the uppermost tier. Each occupier
of each lab bench faced us and smiled in invitation. If we knew then what we
know now, we wouldn’t have followed instructions so complacently. We would have
shifted our levitation fields or mumbled amongst ourselves, hoping for more
information. Maybe at least one of us would have begged to differ. But we
didn’t know anything, so we stood, hundreds of us bunched into the near half of
the room, levitating and clicking away inside ourselves, waiting for
instruction.
oOo
An ebook version (pdf, mobi, lit, lrf,html) of We, Robots is available from Book View Cafe.
Published by Book View Café, Cover design by Deb Deysher (http://www.doubledmedia.net/portfolio.htm)
We, Robots A Novella by Sue Lange was originally published
in January 2007 by Aqueduct Press as Volume 16 in the Aqueduct Press
Conversation Pieces Series.
