Information Technology Reference
In-Depth Information
suggest that the Turing Test is too restrictive, whereas others sug
gest it is too inclusive. Although the Turing Test has its limitations,
other testing methodologies also may have shortcomings.
Some current computer systems demonstrate intelligence within
a specific area of application. In some areas, computer systems per
form as well or better than human experts. The methods used in
these computer applications typically take advantage of qualities of
computers and may be quite different than methods that humans
would use. To date, these application areas are narrow in scope,
and modern computers do not demonstrate the breadth of knowl
edge that we associate with general intelligence.
Woody Bledsoe, a pioneer in the field of artificial intelligence,
articulated a vision that one day computers will be able to demon
strate intelligence, acting as a friend and advisor to people, and
many researchers in the field today share that vision. Current sys
tems are far from attaining that general vision, and it remains to be
seen whether that vision can be obtained.
■
Terminology from This Chapter
blackbox testing
Turing Test
whitebox testing
■
Discussion Questions
1.
The Turing Test is often cited as one mechanism for deter
mining whether an entity is intelligent.
a.
If you needed to update this test to take advantage of mod
ern technology, how would you propose that a current ver
sion of the test be run?
b.
To what extent does this test seem adequate or appropri
ate?
2.
The chapter outlines both blackbox and whitebox testing to
determine whether an entity is intelligent. Organize a debate
in the class, with one group arguing each of the following po
sitions:
a.
Blackbox testing is the appropriate approach for testing
intelligence.
b.
Whitebox testing is the appropriate approach for testing
intelligence.
