INFORMATION TECHNOLOGY SYMPOSIUM
19 7
won't find the boundaries of their capability any more quickly than we
find them in other humans. So when they claim to be human, when they
claim to be conscious, they will be very convincing.
In
fact, they'll be
very intelligent, so they'll succeed in convincing at least many of us that
they are human and that they are conscious. But some philosophers will
say, "No, you can't be conscious unless you squirt neurotransmitters or
unless you are based on DNA-guided protein synthesis." We have a sim–
ilar debate today with animals. People say, "Well, okay, the behavior of
certain advanced animals seems like human behavior in certain ways,
but they don't have advanced language capabilities like humans, so
they're not conscious, they are just these sort of automatic machines."
In
my view, fundamentally there's no absolute scientific objective test
for consciousness. We can test for correlates of consciousness, behav–
ioral correlates, but such a test would be based on a fundamental
assumption that these human entities are conscious. We all assume that
other humans who at least appear to be conscious are conscious. I
assume that all of you are conscious; if I keep speaking much longer,
maybe you won't be, but outside of our shared human experience, if we
look at the potential consciousness of animals, shared human consensus
breaks down. This will become even more controversial when we have
non-biological entities that claim to be conscious. However, these non–
biological entities will appear to be human to a greater extent than ani–
mals appear to be human because animals really are distinctly different
in their capabilities.
If
we have perfectly replicated a human brain-and
this is a technical task that is not impossible-it's quite feasible that
machines will appear to be human. We can touch and feel all the tech–
niques needed to do that today; we can't quite do it yet however-we
don't have the miniaturization; we don't have the computational
power-but if you look at the current trends, we will be there within a
few decades. These will become feasible scenarios.
I want to finish with one more scenario, and I'm sure other issues will
come up later. How will all this affect human civilization? I don't view it
as an invasion of intelligent entities coming over the horizon.
In
my opin–
ion, we're going to enhance our own intelligence and our own experience
by becoming very intimate with our technology. We're very intimate with
computer technology today; it's very much integrated into our human
civilization, into our human decision-making. A significant fraction of
financial decisions, for example, are made by computer systems, by
neural nets and evolutionary algorithms, which are simulations of evolu–
tion in the computer; the results are chaotic and unpredictable. Comput–
ers are making decisions in our financial markets and pattern-recognition
