Transcript How to Think about the Turing Testx

How to Think about
The Turing Test
A Philosopher’s Perspective
A Question
• Why the strange set-up of the Turing-Test?
Why did Turing ‘pit’ a machine against a
human in some kind of contest?
Why not have the interrogator simply interact
with a machine and judge whether or not the
machine is intelligent based on those
interactions?
The Super-Simplified Turing Test
Interrogator
Machine
Obvious Answer: Bias
• The mere knowledge that we are dealing with
a machine will bias our judgment as to
whether that machine can think or not, as we
may bring certain preconceptions about
machines to the table.
• Moreover, knowing that we are dealing with a
machine will most likely lead us to raise the
bar for intelligence: it can’t write a sonnet?
Ha, I knew it!
• By shielding the interrogator from the
interrogated, such a bias and bar-raising is
eliminated in the Turing-Test.
The Simplified Turing Test
Interrogator
Machine or Human
Level the Playing Field
• Since we know we might be dealing with a machine,
we still raise the bar for the entity on the other side
being intelligent.
• Through his set-up of the test, Turing made sure that
the bar for being intelligent wouldn’t be raised any
higher for machines than we do for fellow humans.
Still …
• While having the ‘contest-like’ set-up may make sense from
the standpoint of an imitation game played with untrained
judges, as a scientific test, one would hope that any kind of
trained researcher would be able to see past the fact that
they are dealing with something that is not human and not
have that interfere with any ascriptions of intelligence.
“Contrary Views”
• After Turing laid out the Turing Test in his
original paper, Turing goes over a list of
“Contrary Views on the Main Question”:
• Machines:
– can’t make mistakes
– can’t be creative
– can’t learn
– can’t do other than what they’re told
– etc
A Puzzle
• That’s weird!
– Many people regard Turing’s paper as proposing the
Turing Test as some kind of practical test for machine
intelligence.
– But if that were so, you would think that Turing would
address objections of the kind that maybe something
can pass the test without being intelligent.
• Instead, it seems like Turing addresses objections
to the claim that machines can pass the test.
• Why?
My Answer
• I propose that the convoluted set-up wasn’t merely a practical
consideration to eliminate bias in some strange game, but
rather to confront us with the very prejudices that, at Turing’s
time, many people had against machine intelligence.
• Thus, the ‘Turing Test’ isn’t at all meant like a practical test to
deal with our prejudices or biases regarding machine
intelligence, but rather a thought experiment meant to
merely point out these prejudices and biases, and encourage
us to think more rationally about machines and machine
intelligence.
• Indeed, the ‘Objections’ that Turing addresses aren’t so much
objections to machines being able to pass the Turing Test, but
rather objections that go straight to the issue of the possibility
of machine intelligence.
Language
• Another way of looking at the Turing Test is that if we put a label
‘intelligent being’ on other human beings based on their behavior then,
just to be fair, we should do the same for machines, whether we are
correct or precise in any such attributions or not.
• In other words, Turing’s point was that we don’t have a precise definition
of ‘intelligence’, and instead of trying to provide some kind of operational
definition (as some commentators see it), Turing took this as a given, but
whatever fuzzy or sloppy concept of ‘intelligence’ is, we should at least be
consistent in our use of slapping this label onto things (humans or
otherwise).
• Also, as our concepts and theories progress, our scientific investigations
into intelligence will likely go far beyond having a conversation with a
candidate intelligent entity, so at that point we will have other criteria to
consider (but at this time, our judgments regarding intelligence of entities
are still mainly based on interactions with those beings, so Turing can be
excused to point to those interactions).
‘Imitation Game’ vs ‘Turing Test’
• In other words, I think it is likely that Turing
never intended to propose any kind of test for
machine intelligence (let alone propose a
definition!).
– Interesting fact: In his original article Turing uses
the word ‘pass’ or ‘passing’ 0 times, ‘test’ 4 times,
and ‘game’ 37 times.
• It is therefore also likely that commentators
who criticize the Turing Test as a test, or try to
propose better tests (e.g. the Total Turing
Test), are completely missing the point!
The Turing ‘Test’ as Harmful!
• Moreover, I believe that regarding Turing’s contribution as
laying out a test is harmful.
• The harm is that we have been thinking about the goal of AI in
these terms, and that has been, and still is, detrimental to the
field of AI.
• E.g. In “Essentials of Artificial Intelligence”, Ginsberg defines
AI as “the enterprise of constructing a physical symbol system
that can reliably pass the Turing Test”
• But trying to pass the test encourages building cheap tricks to
convince the interrogator, which is exactly what we have seen
with Eliza, Parry, Cleverbot, Eugene Goostman, and pretty
much any entry in the modern-day versions of the Turing Test.
• This kind of work has advanced the field of AI, and our
understanding of intelligence … exactly zilch!
Grand Challenges
• OK, but maybe the Turing Test (and the Loebner
competition) is a kind of Grand Challenge?
–
–
–
–
Landing people on moon
Chess (Deep Blue)
Urban Challenge
Jeopardy (Watson)
• But at this point in time, I feel that trying to
create human-level intelligence in a computer is a
ridiculously-grand challenge, and hence a bad
Grand Challenge
How to Read Turing’s Paper
• So what did Turing himself really mean? Taken literally, this is an issue of
history, not philosophy. And frankly, I am a philosopher, not a historian.
• To me, a better question to ask is: What, if anything, can we learn from
Turing’s paper?
• Well, there are many interesting parts of the paper, especially in Turing’s
responses to the ‘Contrary Views’.
– E.g. his distinction between “errors of functioning” and “errors of conclusion” is an
incredibly important distinction between operation and description that I believe too
many AI researchers do not take seriously.
• But I believe the most important reading of his paper is to see the Turing
‘Test’ as a statement about the use of the word ‘intelligence’.
• That is, rather than an actual, practical, test, I believe we should look at
the Turing Test as a thought experiment that forces us to examine our
preconceptions (and prejudices!) regarding the concept of intelligence.
• In fact, I propose that we no longer refer to the Imitation Game as the
Turing ‘Test’!!
Pluto and Planets
• Asking how many planets there are in our solar system seems
to be a factual matter:
– We believe there is a straightforward fact of the matter to this issue.
• If I say: “There are X planets in our solar system” then this statement is either
true or false.
– Thus: How many planets there are is an empirical issue: observations
will tell us how many there are
• However, as the case of Pluto demonstrated, things aren’t
that easy. Apparently the issue of how planets we have in our
solar system isn’t just an empirical issue, but it is also one of
interpretation.
• Maybe the same is true for machine intelligence!
– Yes, we have to see how machines function and what machines are
able to do … but once we know all about that, we still have to see if we
want to interpret that as being ‘intelligent’.
Artificial Flight and
Artificial Intelligence
• Imagine going back 100 years when the Wright
Brothers had their first flight.
• We can imagine people say: “Well, but that’s not real
flight. There is no flapping of the wings!”
• But over time, we realized that, from the standpoint
of using concepts that help us think, explain, predict,
and otherwise make sense of the world around us, it
is a good idea to consider airplanes as really flying.
• Again, maybe the same is true for intelligence!
The original question, “Can machines think?”, I believe
to be too meaningless to deserve discussion. Nevertheless
I believe that at the end of the century the use of words and
general educated opinion will have altered so much that one
will be able to speak of machines thinking without expecting
to be contradicted.
-Alan Turing (1950)