Introduction to AI - Fifth Lecture

Introduction to AI
5th Lecture - Philosophical
and Ethical Considerations
Wouter Beek
me@wouterbeek.com
6 October 2010

Last week’s ism’s
0 Behaviorism: mental states are attributed based on
external observations.
0 Functionalism: mental states are causal connections
between input and output, i.e. structural
configurations.
0 The study of the brain is irrelevant to the study of the
mind.
0 Biological naturalism: mental states crucially
depend on a neurological substrate.
0 Computationalism:
0 The mind is an information processing system.
0 Thought is computation.

Weak AI || Strong AI
0 Weak AI: machines simulate intelligence / behave as if
they are intelligent.
0 Biological naturalism
0 Strong AI: machines are intelligent.
0 Behaviorism, functionalism, computationalism
0 Most AI researchers don't care… [Russell&Norvig]

Thought experiments
0 A fictitious experiment that gathers intuitions regarding
some problem statement.
0 Plato’s allegory of the cave.
0 “Much of modern physics is built not upon measurement
but on thought experimentation.” [Martin Cohen]
0 Shrödinger’s cat, Maxwell’s demon, Galileo’s Tower of Pisa
experiment (1628)
0 Crucial to philosophy of mind and philosophy of AI.
0 Often related to SF literature: time travel, zombies, strange
machines.

Chinese room experiment
0 John Searle, 1980, Minds, Brains, and Programs
0 A human is inside a room and is handed programs and
data.
0 By following the programs meticulously, the human is
said to ‘translate’ and ‘understand’ his data
manipulation task.
0 Behaviorism, functionalism, computationalism
0 But the human does not understand the manipulation
task at all!
0 “Programs are neither constitutive nor sufficient for
minds.”
0 Thought requires intentionality.

Intentionality
0 The property of mental states to be directed towards
some object, i.e. to be about that object.
0 Intentionality is a characteristic of all and only acts of
consciousness.
0 Thus setting conscious phenomena apart from physical,
unconscious phenomena.
0 According to this definition:
0 No machine can be conscious.
0 Syntactic operations need not be indicative of semantic
content.

Qualia
0 The unit of subjective conscious experience.
0 The way in which things seem to us.
0 The “what it is like”-aspect.
0 For instance:
0 The pain of a headache.
0 The smell of flowers.
0 The red color of tomatoes.
0 Qualia pose a problem to a materialist world-view.
0 But remember: most AI researchers don't care…
0 Could this be a blind spot to AI research?

Related to the “argument
from various disabilities”
0 “Be kind, resourceful, beautiful, friendly, have initiative,
have a sense of humor, tell right from wrong, make
mistakes, fall in love, enjoy strawberries and cream,
make someone fall in love with it, learn from experience,
use words properly, be the subject of its own thought,
have as much diversity of behaviour as a man, do
something really new.” [Turing1950]
0 If qualia are not needed in order to replicate these kinds
of behavior, then an AI researcher couldn’t care less.
0 But if qualia are necessary in order to replicate certain
forms of behavior, then weak and strong AI become the
same undertaking.

Mind-body problem
0 How are mental states related to bodily
states?
0 Materialism: there are no immaterial
aspects of thought.
0 Compatible with functionalism and
strong AI.
0 Cartesian dualism:
0 The immaterial mind and the material
body are ontologically distinct, yet
causally related.
0There is some bit of magic to the brain
that makes it connect with an immaterial
mind.
0Compatible with biological naturalism and
the existence of intentionality and qualia.

Philosophical zombie
0 Like a normal human being, but lacking qualia.
0 When it sees red tomatoes it can ascertain that they
are indeed red, but cannot consciously experience
their redness.
0 Problem of (the existence of) other minds.
0 We presuppose that one can lack qualia and yet still
be a human being in all physical aspects.
0 Thereby presupposing that qualia cannot be
physically motivated.

Mary’s room experiment
0 Mary the scientist lives in a black and white room.
0 She learns all there is to know about the perception of
the color red in physical terms.
0 I.e. a functionalist description of the process.
0 E.g. how certain wavelengths relate to the neurological
state of recognizing something to be red.
0 If Mary leaves the room and observed a red object for
the first time, will she thereby attain new knowledge?
0 Frank Jackson, 1982, Epiphenomenal Qualia.

Leibniz’s mill
“One is obliged to admit that perception and what
depends upon it is inexplicable on mechanical principles,
that is, by figures and motions. In imagining that there
is a machine whose construction would enable it to
think, to sense, and to have perception, one could
conceive it enlarged while retaining the same
proportions, so that one could enter into it, just like into
a windmill. Supposing this, one should, when visiting
within it, find only parts pushing one another, and never
anything by which to explain a perception. Thus it is in
the simple substance, and not in the composite or in the
machine, that one must look for perception.”
[Leibniz, 1714, Monadology]

Brain prosthesis experiment
0 Piecemeal replacement of neurological configurations
by structurally identical electronic configurations.
0 External behavior must stay the same, but the internal
experience goes away.
0 Under the assumption that external behavior remains
unaffected, the waning of internal experience must
proceed at once.
0 This means that any prosthesis, however small, could
result in an instantaneous and complete removal of
internal experience.

Brain in a vat
0 Not about the
functionalism/naturalism-
dichotomy.
0 Because supposed qualia can still
be experienced and attributed to
the neural substrate.
0 It questions the veracity of the
thoughts one entertains.
0 Propositions that relate to bodily
experience are all falsely
entertained.
0 E.g. “I am walking.”

Brainstorm machine
0 Based on the 1983 film Brainstorm.
0 A helmet allows sensations to be carried over from one
person to another.
0 With eyes closed I accurately report everything you are
looking at. I marvel at how the sky is yellow, the grass red.
0 Suppose inverting the connection makes me report the sky is
blue, the grass green. Which is the right way of connecting?
0 Dependent on a calibration of the two subjects' reports.
0 Conclusion: no intersubjective comparison of qualia is
possible. (Remember: the problem of other minds.)
0 Daniel Dennet, 1997, Quining Qualia

Introduction to AI - Fifth Lecture

Technological singularity
0 Machines that surpass human intelligence.
0 Exclusively quantitative view of AI research.
0 ‘Intelligence’ is a word that we attribute to specific
kinds of behavior.
0 Is intelligence an inherently anthropomorphic
attribute?
0 And if it is not, what would it matter for humans to be
confronted with something they cannot understand?

Ethical questions
0 People lose their jobs due to AI.
0 R&N: AI has created more jobs than it has eliminated.
0 But the jobs that are eliminated and created are not the
same. AI catalyzes the class-distinction between high
and low educated.
0 People have too much / too little leisure time.
0 People loose their sense of being unique.
0 R&N: As with Copernicus, Kant, Darwin.
0 But AI not only attacks the ideology of human
superiority, but actively proposes an alternative.
0 People loose their privacy.
0 Loss of accountability.

Introduction to AI - Fifth Lecture

More Related Content

Viewers also liked (20)

Similar to Introduction to AI - Fifth Lecture (20)

Recently uploaded (20)

Introduction to AI - Fifth Lecture

Editor's Notes