296 | 27 THE TURING TEST—fROm EVERy ANGlE
abilities required to pass the Turing test may not be (as optimists like Kurzweil claim) only a
short step from superhuman-level AI, but they could have an enormous impact on human life.
So, why is the Turing test important? The many people in AI who do regard human-level AI
as important and achievable need a criterion to determine when they reach that goal. Turing’s
test provides this. It is also anthropomorphism-proofed. In 1949 Turing’s colleague at the
University of Manchester, the neurosurgeon Geoffrey Jefferson, said:^35
We have had a hard task to dissuade man from reading qualities of human mind into animals. I
see a new and greater danger threatening—that of anthropomorphizing the machine. When we
hear it said that wireless valves think, we may despair of language.
AI researchers often encourage us to treat machines as humans, for example by building hyper-
realistic androids such as Hiroshi Ishiguro’s Geminoids—robots designed as exact copies of
Ishiguro and others, which can be encountered in a public cafe. We require some way of ensur-
ing that judgements of intelligence in machines are not solely the product of anthropomorphiz-
ing. Turing’s imitation game includes a disincentive to anthropomorphize: to avoid getting egg
on their faces, interrogators are abnormally suspicious of the contestants (as in the Loebner
Contests). And even if an interrogator does anthropomorphize the machine contestant, the
three-player game is a blind controlled trial—the interrogator will also anthropomorphize the
human contestant, with the result that there is no unfair advantage to the machine. Whether
by design or accident, Turing’s test addresses the anthropomorphism danger. His test matters.
Smart moves
If a computer beats a human player in a game of chess, does that show that the computer under-
stands chess—or is how the machine plays important? In 2003 Garry Kasparov said that ‘com-
puter superiority over humans in chess had always been just a matter of time’. But, he added,
Deep Blue and similar computers were not what the early chess programmers hoped for ‘when
they dreamed of creating a machine to defeat the world chess champion’:
Instead of a computer that thought and played chess like a human, with human creativity and
intuition, they got one that played like a machine, systematically evaluating 200 million possible
moves on the chess board per second and winning with brute number-crunching force . . . Deep
Blue was only intelligent the way your programmable alarm clock is intelligent. Not that losing to
a $10 million alarm clock made me feel any better.
Others agreed. According to the Economist, Deep Blue showed that ‘chess-playing skill does
not, in fact, equal intelligence . . . it is possible to get a dumb machine to do it better than any
human. The equation of chess-playing with intelligence is centuries old, but it is time to lay it
to rest’. Likewise, cognitive scientist Douglas Hofstadter said, ‘My God, I used to think chess
required thought. Now, I realize it doesn’t’.^36
IBM acknowledged that Deep Blue is ‘stunningly effective at solving chess problems, but
it is less “intelligent” than the stupidest person. It doesn’t think, it reacts’. (For the objection
to Turing’s game that a brute force machine like Deep Blue could fool the interrogator, see
Chapter 25.) The company moved on to new machines, including Watson, and new human–
machine games. According to IBM, Watson can think. In 2011 the computer beat the two
highest-ranked human contestants in an episode of the television game show Jeopardy!