Title: elucidate weaknesses within the philosophy of artificial intelligence AI in an attempt to work towar
1elucidate weaknesses within the philosophy of
artificial intelligence (AI) in an attempt to
work towards a more holistic understanding about
what kind of knowledge satisfies the conditions
of intelligence
abstract
Suitss notion of gaming playing as a basis for
re-defining the criterion of intelligence to
which AI is evaluated
2perhaps this has nothing to do with chess, but
is a game to the death between two forms of
intelligence the bounded field of postmodern
intelligence and the unbounded vectors and
virtualities of digital intelligence. A game of
symbolic exchange in which Kasparov as the last
chess player needs not only to be defeated, but
humiliated by his own imaginary catastrophe. And
he was. (Cook Kroker, 1997).
3Can a machine think as you and I do? Can a
machine have consciousness, feelings, or
knowledge, rather than just simulate these
capacities?
Hard-Ware
Are the kinds of capabilities demonstrated by
machines an indication of intelligence and, if
so, is this kind of intelligence similar to that
of sentient beings, such as humans?
Wet-Ware
4Hard-Ware
Soul
speciesism
Physics
Wet-Ware
5Games such as chess, go, and drafts (or checkers)
are ideal games for computerization. This is
partly because they are games of perfect
information as all the pieces are visible to both
players and there is no element of chance or
probability. This means that it should be
possible to calculate the ideal move for any
given situation, and hence, in theory at least, a
computer should be able to play perfect chess
(pp.108-9).
Is ones capacity to play chess a suitable
indication of intelligence?
6Other sorts of games?
Bernard Suits
Deep Blue
Is ones capacity to play chess a suitable
indication of intelligence?
7More on AI
Imitation Game (Turing Test)
Stelarc Prosthetic Head
Chinese room (Searle), need semantic
8More on AI
Imitation Game (Turing Test)
- What is the relationship between
- appearance of understanding
- understanding
- intelligence
Chinese room (Searle), need semantic
9Imitation Game (Turing Test)
Strong-AI
Weak-AI
Chinese room (Searle), need semantic
10Strong-AI
Weak-AI
Intelligence entirely algorithmic and possible to
formalise through programming
Intelligence is comprised partly/largely by
semantic characteristics, not accessible to
computing. Computers only simulate
11Why Chess?
Human vs. automaton Easily represented by
programming Well-defined rules Easily represented
visually No financial or ethical burden
12(No Transcript)
13Kasparov
sensed a new kind of intelligence
(Deep Blue analyses more than 100 million
positions per second)
14Chess is too easy (Bringsjord 1998)
Deep Blue functions sequentially Human learns -
choose from competing options (intuition)
can a machine tell a story?
Narrative, character adoption
The test give the computer a beginning sentence
15Towards Bottom-Up design
- Not control of nature
- but symbiosis and emergent intelligence
No direct human comparator
16Evolving grasshoppers
Game playing and broad intelligence
Huizinga
Freedom and awareness of non-real status Sense of
disinterestedness (but not diminishing the value)
Suits
Agency, attituderules of higher priority
17Suits
Agency, attituderules of higher priority
18Chess, the wrong kind of game
Sport-like games, non-linear, dis-connected
decision making
Can a machine play sport?
19distributed intelligence (hive mind) Stelarc
again