FoundationalConceptsNeuroscience

players. Consider the quiz game of Jeopardy. Here, too, artificial intel-
ligence has come a long way, and in 2011 a computer beat champion
human players of the game. A Jeopardy-playing computer may still
make seemingly silly mistakes, based on quirkiness in language recog-
nition and interpretation, but the capacity to avoid such mistakes will
continue to improve indefinitely. Machines can be programmed to
create music and art, and the generation of novel ideas by computers
is likely to become increasingly sophisticated with time.
And all this has been achieved with still essentially first-generation
computers, based on ideas of computation described by Alan Turing
(1912-1954) at the dawn of modern thinking about computability
and instantiated in computing machines first built in the 1940s and
1950s. Speed and capacity have grown a billionfold since then, but the
principles are essentially the same.
Some say that the intelligence of computers—their ability to solve
problems, and even generate novelty—will in the near future exceed
that of us humans. This state of affairs has been called the technologi-
cal singularity. What would be the implications of such a singularity?
Is it likely to come with a bang, or a series of bangs, or with a series of
whispers?
The whispers, or bangs, are already happening, in selected
domains: chess, jeopardy, financial markets. It has been suggested
that some responsibility for the economic meltdown of 2008 may jus-
tifiably be placed upon computer systems that were running trading
algorithms beyond the comprehensive understanding and ready con-
trol of humans. The machines were literally out of control. And we are
only at the very beginning of things like this. Who’s to say what the
future may hold? Certainly it will be interesting.
In all existing cases, and many future extrapolations—such as
singularity scenarios—computers still lack minds. No matter how