NewPhilosopherRUNNER-UPA failure of collective
intelligence
by Warwick SmithHumans, individually, can be in-
credibly brilliant, but collectively we’re
often puzzlingly stupid. To take a sim-
ple, uncontroversial example, we know
that the tropical forests, the Amazon
and the forests of South-East Asia, are
the lungs of the world. Despite essen-
tially universal agreement about this
fact, we’re continuing to destroy them
at an extraordinary rate in order to pro-
duce soy beans, beef, palm oil, and more.
It seems that despite our intelligence
and our astonishing global communi-
cation infrastructure, we’re collectively
incapable of aligning what we know
needs to happen with what we do.
There is a thought experiment
about artificial intelligence known as
‘the paperclip maximiser’ – bear with
me a moment, this is related to human
sustainability and regeneration. In this
thought experiment we imagine that
there’s an AI system used by a company
that makes paperclips. This AI is tasked
with increasing paperclip production
and is fed all the necessary information
regarding paperclip making, includingmaterials, labour, human motivation,
supply chains, etc. Critically, it’s also
capable of learning how to learn and
does this at an exponential rate.
To cut a long story short, the AI
gets better and faster at making pa-
perclips at a rate that far exceeds hu-
man capacity to keep up or to adapt.
Eventually it turns the entire universe
into paperclips, with all humans and
the biosphere being consumed quite
early in the process. The thought ex-
periment, not meant to be taken too
literally, was designed to explain that
AI doesn’t need to have general in-
telligence and self-awareness to be a
threat but can be a threat simply by
being single-minded and able to adapt
faster than we can react.
I recently heard Daniel Schmach-
tenberger taking this thought experi-
ment in a very interesting direction
by saying that human society is al-
ready the paperclip maximiser but
instead of making paperclips we’re
making dollars – which are primar-
ily just zeros and ones in bank data-bases. Our collective intelligence sys-
tem has one overriding purpose: to
turn everything into money – trees,
labour, water... everything. It is also
very good at learning how to learn
and is extremely good at eliminating
any threats.
The paperclip maximiser is not
controlled or driven by anybody, it’s
just a product of our global corporate
capitalist economic model. The rules
of the game are such that if you don’t
strive to turn everything into dollars
then you’ll likely be defeated (in what-
ever you’re trying to do) by somebody
or some institution that is.
There are plenty of people, organi-
sations, and political parties who are
desperately trying to assert that some
things are more important than mon-
ey. Most people agree with them, but
the money-making paperclip maxim-
iser continues on regardless, barely di-
verted at all by even the most strident,
most well-supported initiatives (such
as reducing tropical deforestation or
reducing greenhouse gas emissions).WRITERS’ AWARD XXIII:
BEING HUMAN
Writers’ Award XXIII: being human