Robotics Reference
In-Depth Information
lions of such operations, the being ceased to have a phenomenology.
6
It
seems almost inconceivable that we could defend this position,
ergo
we
must accept that B is the truth—the robot has consciousness.
Perhaps this approach, refuting the arguments of the non-believers,
is more defensive than is necessary. A far more positive stance on the
subject was taken by Sam Lehman-Wilzig in his 1981 essay: “Franken-
stein Unbound: Toward a Legal Definition of Artificial Intelligence”.
Lehman-Wilzig presents weighty evidence that robots would be, by most
definitions, alive, citing seven significant achievements of AI that had
already been realised (as of 1981), or were theoretically possible at that
time. These were that robots could
1. Imitate the behavior of any other machine. (This is self-evident,
since software can simulate the behaviour of anything.)
2. Exhibit curiosity. (Examples include search-and-rescue robots.
7
)
3. Recognize and react to the sight of themselves and members of
their own machine species. (In a crude way this was first achieved
in the late 1940s by Grey Walter's tortoises,
8
and is almost trivial
by the standards of today's recognition technology,
9
as is program-
ming robots to react to what they recognize.)
4. Learn from their own mistakes. (Also trivial nowadays.
10
)
5. Be as creative and purposeful as are humans, even to the extent of
looking for purposes that they can fulfil. (The whole of Chapter 5
describes the creativity of robots.
11
)
6
A description, history or explanation of phenomena.
7
See Chapter 8.
8
See the section “Robot Tortoises” in Chapter 2.
9
See the sections “Human Face Recognition” and “Recognition in Three Dimensions”, both in
Chapter 4.
10
See the section “How Computers Learn” in Chapter 6.
11
See also the section “How Computers Discover and Invent” in Chapter 6.
