We were smart enough to build them, but are we smart enough to fight them?Robopocalypseauthor Daniel H. Wilson considers a hyper-realistic future in which machines run amock.
Machine minds are stalking the modern world, carefully changing our
lives in myriad subtle ways. Although they are largely invisible, powerful
artificial intelligence (AI) algorithms support the most basic workings of commerce
and warfare in our society. They advertise products to us, build them, and then
ship them out. AI is employed to spy on our enemies, plan the logistics of our
wars, and then to fight them. And as if these profound undertakings weren’t
enough, the machines can also beat us senseless at Jeopardy.
It’s getting tough out there for a human being.
My novels Robopocalypse
and its soon-to-be-released sequel, Robogenesis,depict a terrifying, hyper-realistic
future in which these thinking technologies have run amuck and then evolved.
And the scariest part is that it’s all based on real-world research in robotics.
Before writing fiction, I studied robotics at Carnegie Mellon
University. My first week as a student in the Machine Learning Department, I
learned of a project that could predict flu outbreaks by modeling how much orange
juice people were buying in real-time (from loyalty card data). That’s a little
creepy, but what floored me was that the same
loyalty card data also helps determine which products are placed on shelves,
where the shelves are placed in the store, and where the stores are built in
the cities. The very infrastructure of our cities is covered in alien
I realized then that this is truly a machine’s world--we’re just
living in it.
Within a few years, my own research was using complex models of how
people behaved to predict and influence their actions. My thesis was called
“Assistive Intelligent Environments for Automatic Health Monitoring” and it used
AI for the noble cause of helping older people to live safely and independently
in their own homes for as long as possible. I designed algorithms to perform
real-time tracking and activity recognition because I grew up watching my
mother (a nurse) struggle to manage the decline of her elderly patients who
were scattered across rural Oklahoma.
To keep these vulnerable people safe, I used sensors in the home to learn
normal patterns of behavior. Then, any deviation in the long-term pattern
(usually associated with functional decline) could be spotted immediately. When
speaking to a nurse, patients will often avoid the truth out of embarrassment
or shame or plain old confusion. But it was impossible to elude my algorithms
once they recognized the change in pattern created when a person stops being
able to bathe, or clean house, or safely cook. Like so many machines, mine was
designed to take care of you whether you liked it or not.
It’s enough to get a novelist to thinking.
Anywhere you care to look, a machine intelligence is likely to be
hiding just out of sight, watching us, modeling our behavior, and pulling
hidden strings. Sometimes they work for the good of all humankind, and
sometimes for the good of a corporate or political entity: cameras track our
gaze and paths through department stores; automated face recognition algorithms
sift through Facebook to learn the nitty gritty architecture of our social
networks; and data-crunching AI algorithms are credited with being instrumental
in the re-election campaign of our current president.
As a society, we are deliberately teaching the machines to know us--literally better than we know
ourselves. With near-limitless processing and storage capabilities, the
machines can pay attention to the minute details of our lives over extended time
periods. Often this is a good thing, as machines track our health and grow our
food and streamline our products.
But as a writer, I’m drawn to thinking about a more dramatic scenario:
What happens if we go to war against a God in a box?
This is what my new novel Robogenesis
contemplates. How could we as a species survive a fight with thinking machines who
are far smarter and stranger than we can imagine? Coming from the laboratory, I
believe that the most terrifying mechanical adversaries are not the ones chasing
us with shotguns; they are the quiet ones--the hidden, patient ones--robots who
are much, much smarter than we are, and who are lurking behind every corner of
our modern world, watching and learning more about us every day.
We were smart enough to build them, but are we smart enough to fight