Robotics Reference
In-Depth Information
AndIcanevenseeicebergsdowninthefjord.”Well,thatwasit. So
much for the mass raid of Soviet nuclear ballistic missiles. It was just
the moon rising over Norway and headed towards Greenland. Nobody
in the warning system's software design team had thought about this
possibility!
Moor's question, “Are there decisions computers should not make?”,
depends for its answer on what kinds of decisions computers can and
cannot make competently. The problem is that, although well-written
software will normally outperform skilled humans at many of the most
challenging intellectual tasks, sometimes the software will not return the
correct answer, for whatever reason. In October 1960 the world was
extremely fortunate that someone in Colorado had the good sense to ask
what could be seen through the window of a building in Greenland, and
there have been other, equally frightening examples since then.
A Code of Ethics for Robots
In the 1920s science fiction first became popular. One of the stock plots
involved the invention of a robot that ultimately destroyed its creator,
Frankenstein
and
Rossum's Universal Robots
being the best known exam-
ples of this genre. In 1942, tired of the repeated use of this particular
plot, Isaac Asimov published a short story called “Run-around” in which
he stated three laws of robotics:
1. A robot may not injure a human being, or, through inaction, allow
a human being to come to harm.
2. A robot must obey the orders given it by human beings except
where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection
does not conflict with the First or Second Law.
Later Asimov added a further law, protecting the whole of humanity at a
stroke:
4. A robot may not injure humanity, or, through inaction, allow hu-
manity to come to harm.
Although Asimov's laws were created for science fiction, they are valu-
able in developing an ethics for the world of true science. But are they
enough? Specifying what robots must
not
do is a start, but it is hardly
