by Eileen Neary
Assistant Project Manager
Ever since the creation of
automated machines, the fears of an artificial intelligence (AI) takeover (or cybernetic
revolt, as more professional futurists call it) have been growing. From the
minds of science fiction greats Isaac Asimov and Aldous Huxley to 2015’s
blockbuster films Avengers: Age of Ultron
and Ex Machina, culture shows that
people are fascinated with what machines could someday be. This potential
future has long seemed fictional, but the (small) possibility of artificial
intelligence destroying the human race has, in some ways, been on the minds of
more than just storytellers and conspiracy theorists.
The phrase “artificial intelligence”
was coined by cognitive scientist John McCarthy in 1955. By the early ’70s, one
of the first successful QA, or question-answering, programs was functioning. It
was named SHRDLU, and
could follow a user’s instructions to pick up and place blocks and pyramids in
a virtual toy box. Today, “intelligent personal assistants,” otherwise known as
AI software agents, are standard features on the biggest brand-name products.
Perhaps what started with SHRDLU, has continued with agents like Apple’s Siri,
Microsoft’s Cortana and Amazon’s Alexa . . . but where does it end?
Maybe with doom. An informal survey
of participants at the Global Catastrophic Risk Conference published by the University
of Oxford’s Future of Humanity Institute (FHI) estimated in 2008 [PDF link]
that the risk for Superintelligent AI resulting in the end of the human race
has a 5 percent chance of occurring before the year 2100 (a tie with molecular
nanotechnology weapons). This beats out the survey responders’ estimated
probabilities for wars (4 percent), engineered pandemics (2 percent), nuclear
war specifically (1 percent) and other possibilities, like nanotechnology
accidents and natural pandemics.
High-profile scientists like
Stephen Hawking, Elon Musk of SpaceX and Steve Wozniak of Apple, have signed an
open
letter from the Future of Life Institute (an organization with the goal to
prevent any existential risks to the human race) that addresses how to make AI
safe and helpful to society, rather than more powerful and potentially deadly.
The letter advocates for the ban of autonomous weapons to prevent a “global AI
arms race.” The goal of the letter is to prevent automated weapons of the
future for a variety of ethical reasons, as the results could be tragic for the
human race.
There has also been the fear of machines
continuing to put employees out of work. But as it turns out, there are some
tasks that robots just cannot do. Amazon
Mechanical Turk is just one host for HITs, or human intelligence tasks (tasks
that require a person's mind). These types of jobs involve taking surveys,
describing and labeling the content in images or videos, matching data, and
conducting research.
If you’re still feeling a little
uneasy about all of this, here’s a list of 10
jobs that have a less than 1 percent chance of ever being replaced by robots,
according to researchers at the University of Oxford (yes, the same university
with an institute that surveys estimates for the probability of human demise.)
And if you’re still feeling uneasy,
well, here’s a cute
puppy.
Did You Know?
Besides Mechanical Turk, other
sites to perform HITs include ShortTask, CrowdFlower and Clickworker. Certain
apps are available too, like EasyShift and TaskRabbit.
No comments:
Post a Comment