Or: One Day My Artificial Prince Will Come
Season 3 of Star Trek: The Original Series ended one month before the United States used ten thousand lines of code to land a man on the Moon. In that year, humanity’s perception of space evolved from a terrifying no man’s land to the “Final Frontier.”
Today, the TV series Westworld engrosses millions with sentient androids who walk among us, but only to the extent that we allow them. We think of them as physical extensions of our hopes and dreams. Faster. Stronger. Better. (Probably harder, too.) How far away are we from reaching a new “final” frontier? Or, a better question: how far are they from reaching us?
There’s a lot to glean from pop culture’s revelations about society’s thoughts (and fears.) A year after production on Westworld began and British “psycho-techno thriller” Ex Machina won an Academy Award, Dr. Martine Rothblatt wrote the book on digital immortality. Virtually Human is not a light read; it’s 307 pages of dry neuroscience, software theory, and technological philosophy.
Ex Machina, Westworld, Black Mirror, and others are introducing AI to a wider audience. But how close are we to making AI a commodity?
Westworld Hosts pass the Turing test
No matter how many times an Apple pundit calls the latest iPhone “magical,” Siri still isn’t magical enough to ace the Turing test by which we determine if a computer has true artificial intelligence. Aleksandar Todorović monitors Turing test attempts and results on his website. Here’s how he describes the process:
It’s simple. Put a computer (A) and a human (B) on one side and a human tester (C) on the other side. If the tester (C) can’t recognize which candidate is human and which candidate is a computer after a series of questions, then the computer successfully passed the Turing test.
Some say that, after a machine reaches a certain point in intelligence, that machine would be regarded as “conscious.” In her book, Dr. Rothblatt describes what she calls cyberconsciousness (which the rest of us call “artificial consciousness”) as “a continuum of software-based human-level autonomy and empathy as determined by consensus of a small group of experts in matters of human consciousness.” TL;DR: a Turing test, by committee.
Plenty of developers claim their chatbots passed the Turing test, but none actually have without cheating. Computer scientist Ray Kurzweil thinks a computer could pass as soon as 2029, just in time for late-blooming millennials to start marrying their Persocoms.1
The bridge between social media avatars and Westworld Hosts
If you’ve spent any time lurking on Twitter, you know a bot when you see it. Bots use collections of pre-programmed phrases to cobble together tweets. Sometimes they’re capable of responding to real people with full sentences, but they are by no means intelligent, let alone sentient.
Some Twitter managers use bots to quickly respond to customer complaints from anywhere in the world. Engineers use machine learning to teach them to read naturally-written sentences, select from a list of best possible suggestions, and reply back in plain English. The more real people they communicate with, the more refined and natural their speech becomes. (Think of Summer Wars’ “Love Machine,” minus the nuclear codes.)
Twitter bots are a rudimentary version of Westworld Host programming; the faster our machines can learn from us, the closer we’ll come to having full, intelligent conversations with our computers.
To create a machine capable of transporting hundreds of humans by air, we didn’t build a bird. We invented a plane. To create a machine capable of humanity, we’re not going to build a human. Just like we programmed smartphones to help us avoid traffic and translate Greek to French, we’re going to hardwire an android to think, learn, and feel like a human.
Avatars built from big data and machine learning2
But an intelligent machine does not a human make. Not yet finished tossing 90s prefixes around like frisbees, Dr. Rothblatt coined “mindware” to describe the online real estate on which we pitch our digital tents.
Almost every day we post photos of our family and friends on Facebook; share our favorite things on Pinterest; tweet about our days on Twitter; and search for items we want on Amazon. Our credit cards leave a paper trail of our interests across the web, allowing businesses to monitor our buying habits and serve us customized ads that both intrigue and frighten us.
These bits and pieces of our identities are already being used today to create a user profile. Marketing nerds call it behavioral analytics. (Sound like a conspiracy theory to you? Big-box brands like Amazon and Target have been using these analytics for half a decade.) As smaller, more powerful hardware becomes available, small businesses—and later, the consumers themselves—will be free to improve upon machine learning implementations.
Teaching a robot to love
Teaching a robot to learn is one thing, but many scientists have straight-up rejected the concept of teaching a robot to feel.
That said, scientists like William Sims Bainbridge are already working on hand-coding personality attribute software, using statistics to increase or decrease the probably of certain traits.
And teaching a robot to love? They’re working on it. Dr. Rothblatt boiled down the science of coding emotions in Virtually Human.
During this stage in AI’s development, it isn’t difficult to train machines to recognize the color red. But training it to love—to “consider another subject’s happiness as essential to you own,” as Rothblatt describes—is a project. Human grow into their preferences, but engineers would need to manually program millions of sensory neurons to assign these to the robot.3
Personality, preferences, and love
We can already train machines to react to a subject’s “shapes, colors, scents, phonetics, and/or textures.” Why couldn’t we also trigger positive sensory reactions that become stronger the more you interact with the subject? Wouldn’t we call this chain reaction of senses something like “love?”
Once we’re finally able to code intelligence, apperception, wisdom, and the building blocks of love, we’re one really good 3D printer away from building our own Maeve or Teddy.
And I, for one, welcome our new android overlords.
If you like what you just read and want to delve into some AI-related, cyberpunk awesomeness, please use any of the affiliate links below to try some of my favorite picks on Amazon. Buying from my affiliate link helps to support my work!