Decolonizing Artificial Intelligence
[Music]
[Music]
ai
or artificial intelligence is routinely
presented to us as a monolithic
technology
clean clear and game-changing people
talk about the era of ai
and ai-powered systems and worlds it
feels inescapable
a future that we are marching toward or
already in
so why would other stories about ai be
valuable or useful
well i believe that we need to tell
other stories about ai’s past
its present and even its future stories
that reveal other truths and other
possibilities
and make room for a more sustainable
safe and responsible ai
and ultimately a more human-centric one
so where would you start
well you’d start by looking at ai’s
founding moments
an attempt will be made to find how to
make machines use language
form abstractions and concepts solve
kinds of problems now reserved for
humans and improve themselves
those words frame a research proposal
written in 1955
to fund a two-month ten-man study at
dartmouth college the following summer
the authors came from elite american
organizations harvard
mit bell telephone labs ibm research
and they had diverse backgrounds and
interests and they had an ambitious
agenda to make machines think like
humans
the expectation was that computing
technology would continue along the
expansive trajectory established in the
late 1940s and 1950s
the participants imagined a vast array
of computational power and tremendous
possibility
and as a result they believe that much
of that initial research agenda could be
achieved within a decade
this was not to be but the important
thing to remember here is that ai isn’t
new
it’s not a 21st century thing we’ve been
talking about it since 1955.
and in fact whilst the term might have
been coined in 1955 the idea started
even earlier
from 1946 to 1953 the macy’s conference
has convened 10 meetings in new york
city and princeton
bringing together a range of thinkers
from across the disciplinary spectrum to
explore the ideas of the human
machine nature system they called this
cybernetics
curated in part by anthropologist
margaret mead and gregory bateson the
meetings were radically
interdisciplinary and represented an
attempt to constitute a whole new body
of academic knowledge
and a whole new discipline the attendees
are a remarkable role call of big
thinkers
shannon licklider von neumann mcculloch
ashby rosenbluthevena
they were inventing the future and many
of them would go on to be present again
in 1956 in person or in spirit
looking back on it 30 years later me
described it like this
in a sense it was the most interesting
conference i’d ever been in because no
one knew how to manage these things yet
ultimately cybernetics proved to be
contested territory and it was unduly
linked to forms of social engineering
that were unpalatable in 1950s america
and so it’s been conveniently forgotten
inside the ai
story of course there are lots of kinds
of forgetting and silences when it comes
to ai
for as much as we can forget that ai was
a term coined in 1955
we can also forget that the first
conference was funded by the rockefeller
foundation and had participation from
two of america’s biggest companies at
that time
ibm and bell telephone oh yeah and the
rand corporation with its very
complicated relationship to the united
states military was also an active
backer
in focusing just on the technological
piece of ai we sometimes miss the other
bits
where the resources came from why and
what interests were being served by them
we can also ignore that it’s not just
one ai but many built inside many
companies and governments and
non-governmental organizations
across the whole world in different
locations based on different data sets
and different constraints and even
manage through different frameworks and
regulations
and all those different ais they do
different kinds of work
some are about efficiencies and
productivities others are about safety
and control
and yet others are about surveillance
and desire
in 1956 the original ai proposal made
mention of ai and creativity
and the authors speculated about how ai
might make new artistic forms
but that speculation gave way quickly to
more conservative understandings of
intelligence
things about strategy reasoning and
language
yet a little more than 10 years after
dartmouth on the other side of the
atlantic
a remarkable woman installed her first
exhibit
she called it cybernetic serendipity and
it showcased work sitting at the ex
intersections of computing and art it
had music light poetry sculpture
all created with and through computers
jason reichardt was interested in the
ways that randomness would make
art and she imagined that computers
would surely play a part in that
one of the pieces exhibited there was
called return to a square
it was art created using fortran an
early programming language
it was created by a group of artists in
japan calling themselves the computer
technique group
and it was sponsored by ibm research in
japan
that group’s manifesto suggested a
completely different way of framing ai
one that was relational and involves
humans and society never just the
technology
in the manifesto they wrote as follows
we will tame the computer’s appealing
transcendental charm
and restrain it from serving established
power this stance is the way to solve
complicated problems in machine society
those are five stories about ai and they
certainly aren’t the usual ones
each represents a different way into ai
and opens up a different way of seeing
it and thinking about it
all its histories and its contexts the
examples i’ve used here help to reframe
ai and break down some of its powerful
story
we see histories and pre-histories and
silences and erasures we read against
the grain and listen to its hidden
stories
this approach is sometimes called
decolonization
this approach finds its shape in the
social sciences and builds on two really
simple ideas
number one the colonial act or another
act of a powerful force encountering the
world
shapes reality often through a violent
rearrangement of facts bodies and
cultural institutions
it’s also in decolonization possible to
read earlier states and alternative
possibilities inside that same reality
offering up a more complicated and
complicating story
for me decolonizing ai helps open up a
space for all those different kinds of
conversations
and how would such a conversation start
well it might start here
i want to acknowledge the traditional
owners of the land on which i am
standing this is not a wall in nambry
land
never seated always sacred and i pay my
respects to the elders past and present
of this place
i also acknowledge that we are gathering
in many places today and i pay my
respects to the local traditional owners
and elders of all those places too
it means a lot to me to get to say those
words and to dwell on what they mean and
what they signal
and to remember that we live in a
country that has been continuously
occupied for more than 60 000 years
aboriginal people built worlds they
built social systems
they built technologies they built a way
to manage this place and to manage it
remarkably over a protracted period of
time
and every time any one of us stands on a
stage as an australian
here or abroad we carry with us a
privilege and a responsibility because
of that history
and it’s not just a history it’s a
legacy of this place and it should run
through
all our bones and it should be the story
we always tell and the story we always
start with
and that’s a responsibility i take
really seriously and one that shapes all
the work i’m doing here in australia at
the australian national university
here on nanowall and nambry land we’re
building something new
we founded the 3a institute nearly
exactly three years ago
in september 2017 it has one deceptively
simple mission
to establish a new branch of engineering
to take ai safely sustainably and
responsibly to scale
so how do you build a new branch of
engineering in the 21st century
well we’re teaching into existence
through
experimental educational programs we’re
researching into existence with field
sites as diverse as shakespeare’s
birthplace and the great barrier reef
not to mention the world’s largest
autonomous mine and we’re theorizing it
into existence paying attention to the
complexities of cybernetic systems
we’re working to build something new and
something useful
and something that creates the next
generation of critical thinkers and
critical doers
we’re doing this through a richer
understanding of ai’s many past
and many stories and by working
collectively and collaboratively
by teaching and research and engagement
and by focusing as much on framing
questions as problem solving
we’re not building a single ai we’re
making the possibility for many
we’re actively working to decolonize our
imaginations and to build a curriculum
and a pedagogy
that leaves room for many different
conversations and possibilities
we’re making and remaking and i know
will always be a work in progress
[Music]
[Applause]
you