AS ROBOTS BECOME SMARTER AND SELF-AWARE,
SCIENTISTS, THEOLOGIANS CONSIDER THEIR
HUMANITY
BY MARGIE WYLIE
Newhouse News Service
FOR THOUSANDS of years we have
used mythical robots to explore
the question of what makes humans
human.
In the Middle Ages, Jewish
cabalists spun myths about golems,
clay creatures animated by the
secret name of God. The ancient
Greeks sought to create
homunculus, a tiny proto-person
servant. More recently, Mary
Shelley's ``Frankenstein''
creature and the android ``Star
Trek'' crew member Commander Data
have raised the question: ``Can
man-made creatures have souls?''
Anne Foerst's calling is to ask
that question, but not about
mythical creatures. As resident
theologian at the Artificial
Intelligence Laboratory of
Massachusetts Institute of
Technology, Foerst has spent the
past four years pondering how
increasingly smart machines may
affect our sense of humanity.
"I think that computer science,
and especially artificial
intelligence, is the field for
religious inquiry," says Foerst,
a German research scientist who
has served as an ordained minister
and holds a doctorate in theology
as well as degrees in computer
science and philosophy.
In biology or astronomy, the
questions theologians ask deal
with God as a distant and powerful
being. But in the field of
artificial intelligence, the
theological issues are more
"personal," addressing God's
relationship to an individual
being.
A human being asks, "Who am I?
What am I doing here? What's the
meaning of my life?" Foerst says.
"Humans have a very strong sense
of specialness, and these machines
challenge that specialness in
extremely profound ways."
Lab director Rodney Brooks invited
Foerst to work as theological
adviser for a new generation of
smart robots that learn by doing,
just like humans.
One of these is Brooks'
brainchild, Cog, a robot built in
roughly human form except that he
carries his "brain" on his back
in a laptop computer. Cog is
designed to discover and adapt to
the world much the same way a
human baby does.
Traditionally, artificial
intelligences -- such as the
chess-playing IBM computer Big
Blue -- are software applications
primed with vast amounts of data
and then given complex rules for
how to make decisions and for how
to learn to make other decisions.
But such a disembodied
intelligence, Brooks argues,
cannot possibly experience the
world as humans do. Only through
experience as a physical being can
smart robots develop emotions,
which he argues are prerequisite
for a truly intelligent being. So
the aim is for Cog to become
conscious of his body, his
surroundings and, someday, his
"self."
When that happens, asks Foerst,
then what?
"At some point, Cog-like robots
will be part of our community,"
she says. If these robots look
like us, act like us, and are
aware, then shouldn't we welcome
them into the community of
humankind? Should we baptize them?
she asks.
The way theologians answer that
question may shed more light on
how humans treat each other than
how they treat smart robots,
Foerst says.
"We're pretty strict about how we
define humanity," she says. "We
often actually exclude humans from
the human community by saying,
'You are just a Jew or just an
African.'
"Isn't it better to widen up the
criteria of what it means to be
human to include chimps and some
smart robots, so then we avoid the
danger of excluding some people?"
When she isn't asking big
questions about human identity in
a technological age, Foerst also
acts as the lab's gadfly, a role
she clearly relishes. "I make
people aware of their assumptions
about artificial intelligence,''
she says, noting that computer
scientists often fail to recognize
their own mythological or
religious biases and end up
calling them science.
Living forever on the Web.
Some scientists "talk about
downloading their brain contents
into a machine and then
downloading it to the Web in order
to live forever, and they're not
even aware that those things are
faith statements," Foerst says.
"I don't want to deny that it
might be possible at some point to
do that. But I wouldn't say it's
the universal answer for death,
either."
Religious examination isn't always
embraced by the scientific field,
and in the super-rational world of
artificial intelligence, Foerst's
work is especially controversial.
Many scientists in this field fear
that, at best, theology muddies
students' thinking. At worst, it
denies that re-creating the spark
of human intelligence is at all
possible.
In 1997, she created the "God and
Computers" project, a credit
course and lecture series that
explores the links between
religion and artificial
intelligence. It was attacked as
"evangelical" by none other than
Marvin L. Minsky, the MIT
professor who founded the
Artificial Intelligence Lab in
1959.
Minsky, like others at the school,
thinks studying theology is
incompatible with computer
science. "The act of appearing to
take such a subject seriously
makes it look as though our
community regards it as a
respectable contender among
serious theories," Minsky
comments by e-mail. "Like
creationism and other faith-based
doctrines, I suspect it is bad for
young students."
But Brooks, who describes himself
as a scientific rationalist and
"strong atheist," says he can
understand how faith can coexist
with science. "From a scientific
point of view, my kids are bags of
skin full of molecules
interacting, but that's not how I
treat them. I love them. I operate
on two completely different
levels, and I manage to live with
these two different levels.
"I suspect the same can be said
of religious scientists."
Exploration of faith
As computer science bumps against
the limits of rationality, more of
its practitioners are feeling
freer to explore their faith.
Leading computer scientist Donald
Knuth recently wrote a book called
"3:16" in which he examined the
third chapter and 16th verse of
every book of the Christian Bible.
"I thought at first I would be
ridiculed; I had this feeling like
I was coming out of the closet or
something," says Knuth, professor
emeritus for the art of computer
programming at Stanford
University. "I hesitantly
admitted to a few people that I
was working on this book on
weekends but got an unexpectedly
warm reaction."
Knuth says he found that "a lot
of computer scientists have a
God-shaped hole in their hearts.''
This fall, Knuth will present a
series of lectures about his faith
titled "Things a Computer
Scientist Rarely Talks About" as
part of Foerst's annual "God and
Computers" forum at MIT.
Foerst says Minsky is right to be
suspicious. "Some theologians are
very anti-technology," she says.
"The first reaction they always
have is fear: `These robots are
different from us. Humans were
created in the image of God.' They
are not even willing to consider
those questions."
As part of her work, Foerst tries
to educate ministers and
theologians about the science of
artificial intelligence.
Brooks says his "ultimate
megalomaniacal goal" is to build
a robot "that is
indistinguishable from a human --
which I won't do before I die. I
admit that."
But some milestones are already
past.
Today, deaf people can hear with
electronic cochlear implants that
tap directly into a nerve in the
ear. Silicon corneas are in the
works. And these two examples are
just the beginning.
"As we start to connect silicon
to biological material, in living
humans, where is the boundary
between personhood and
machinehood?" Brooks says.
"My kids, if they were so
inclined, would rebel by getting
one of those little metal things
in their tongues. Maybe their kids
will rebel against them by getting
a wireless Internet implant right
into their brains. When do humans
cease to have souls based on
technology?
"Having Anne around was my little
attempt at getting some
consciousness up about some of
these issues."