Robotics Reference
In-Depth Information
Are There Decisions Robots Should Not Make?
Intelligent robots, those with consciousness, will not only possess the ca-
pability to do harm and good, they will also have the capacity to make
harmful and beneficial decisions and recommendations to humans and
to other robots. It is therefore important to consider the question: “Are
there decisions computers should not make?” This is the title of a 1979
essay by philosopher James Moor in which he takes issue with the fol-
lowing statement from Weizenbaum's topic Computer Power and Human
Reason :
Computers can make judicial decisions, computers can make psy-
chiatric judgements. They can flip coins in much more sophisti-
cated ways than can the most patient human being. The point is
that they ought not be given such tasks. They may even be able to
arrive at “correct” decisions in some cases—but always and neces-
sarily on bases no human being
should be willing
to
accept. [17]
Here Weizenbaum has overlooked the ability of expert systems programs
to explain and justify their decisions, 7 an oversight compounded by his
claims that decisions made by computers have bases which “must be in-
appropriate to the context in which the decision is to be made.” Moor
refutes Weizenbaum's argument in a different way, by pointing out that,
in principle, no reason exists why an informed outsider, and a computer
could be an example of such an outsider, cannot be a competent decision
maker:
It is at least conceivable that the computer might give outstanding
justifications for its decisions ranging from detailed legal precedents
to a superb philosophical theory of justice or from instructive clin-
ical observations to an improved theory of mental illness, so that
the competence of the computer in such decision making was con-
sidered to be as good as or better than the competence of human
experts. [18]
Moor cites two important matters of competence that are relevant when
a computer makes an important decision: “What is the nature of the
computer's (alleged) competence?” and “How has this competence been
demonstrated?” Moor is happy for computers to make decisions when
and only when they have proved themselves to be more competent than
7 See the section “How Expert Systems Explain Their Reasoning” in Chapter 6.
Search WWH ::




Custom Search