Centaurs
are the future.
Throughout myth and
history, people have had uneasy relationships with the idea of being
superhuman. Most world religions emphasize a clear distinction between God and
man. Not only is the former more powerful than the latter, but mankind is
inherently unable to achieve a godlike state. In those stories where a human
somehow manages to cross the threshold and achieve supernatural power, it
usually spells disaster.
Advanced artificial
intelligence (AI) technology is not unlike a set of superpowers waiting for
mankind to claim them as its own. AI represents an unknowable and impersonal
“other,” but we already interact with it in the real world. This technology
only gets more capable and powerful each day, so we’re going to reckon with it
one way or another. That’s why I encourage people to remember the centaur.
Centaurs are the
mythological creatures with human head, torso and arms mounted on top of a
horse’s body. They’re the fictional horse-human hybrids from Greek
mythology, and they’re surprisingly useful in conversations about
artificial intelligence. When I talk about “centaurs” in this context, I’m
referring to a people using AI to augment their own cognitive abilities.
There’s a clear hybridization happening here: human intelligence is augmented
by AI, which is then augmented again by human intuition.
Centaurs from mythology
occupy a happy middle ground between god and mankind. They are stronger and
faster than any individual person, yet they retain enough human traits to be
recognizable as people. The same will be true of AI-enabled centaurs in the
near future. They’re ordinary humans, but they’re simultaneously much more than
that.
Here’s how we ought to
manage the rise of the centaur in the AI era:
Human cognition is fundamentally
limited, so we must expand it.
We humans can only retain
so much information at once. We have five senses for interacting with the
world, and we swim in an unsteady sea of emotion. On top of it all, we’re vain
and self-conscious.
Although these
neurological traits make us who we are, we’re completely unstable by comparison
to computers. Artificial intelligence presents us with a way to hedge these
liabilities while transcending our inborn limits. By automating low-level
intellectual work that used to require an entire brain, AI will significantly
expand the abilities of the human mind.
Establish augmentation as
a valuable method for solving new problems.
In my younger life, I
used a computer program to visualize a tesseract, which is
a mathematical abstraction, a four-dimensional shape that doesn’t exist in
human life. This program made it possible for me to project this 4-D image into
two dimensions — I could twist and rotate the shape on my computer screen,
causing it to morph in complicated ways. Even though I could see it on the
computer screen right in front of me, I was completely unable to visualize it
in my mind’s eye.
Humans are
problem-solving animals, but this next dimension is centaur stuff. I leaned on
technology to interact with new information and gained an enhanced
understanding of complicated subject matter. Without that program to help me
visualize a tesseract, I’d be stuck reading the same textbooks over and over
again.
Nowadays people love to
use Google search as the solution to any problem. It’s one of the most
mainstream cognitive augmentations we have. We no longer need to recall
specific information; we just need to be able to find that information when we
need it. As AI research and development continues to rise, we’re going to see
new methods of data entry and retrieval make big waves for how people solve
problems. As we augment our abilities with AI technology, we take a step away
from business as usual and move toward a new normal.
Understand that
domain-specific centaurs will change everything.
We live in a golden era
of “narrow AI,” where algorithms excel at exactly one thing and nothing else. A
narrow AI system might play world-class chess,
accurately tag and label photos or predict your next vacation.
When we use these systems to enhance our own abilities, we become
domain-specific centaurs.
I co-founded an AI
startup earlier in my career that developed narrow solutions for the
manufacturing industry. Someone using our system in their own manufacturing
operation would become a domain-specific centaur, gaining a high-resolution
view into how their machinery performs and how it’s projected to perform into
the future. Our algorithms could catch catastrophic failure months in advance
and alert someone to fix the problem before it became a disaster.
In this specific example,
the human working with our system becomes a domain-specific centaur. His or her
awareness of a machine’s performance was significantly enhanced by having
access to our tools. When the system alerted them to a pending problem, that
domain-specific centaur could take the necessary action to save lives and
money.
Organizations aiming to
automate their processes with AI only need to do a Google search or two to find
businesses developing AI-powered tools for their own industry. Contact them,
meet the people behind the company, and vet the technology. Bring a domain
expert with you to ask meaningful questions deserving thoughtful answers. This
person should be able to confirm that a solution actually uses AI methodology
for its process — it isn’t enough for a system to merely have access to data
and use it. That’s not true AI; it's a conventional rule-based approach that
any iPhone app can implement. These kinds of systems aren’t going to push the
boundaries of previous work in the field.
Conversely, if you're
developing a solution in-house, it's important to ask those same questions find
a way to push those boundaries all the same.
Most spheres of human
interest will make tremendous gains with help from AI technology. This is
especially true when that technology is designed to enhance — instead of
replace — a human approach. Mythology might be full of violent stories from the
past, but its images of the centaur is completely pertinent to modern times.
Credit: https://www.forbes.com/
No comments:
Post a Comment