Most artificial intelligence (AI) platforms hardly merit the name, and none will ever form emotional relationships.
I spent some part of a morning following our new little robot vacuum cleaner around the house. It cleans well enough but it’s not very bright. It chirps annoyingly, like R2-D2, but I can honestly figure out what it needs from the context of the chirps. It is very needy, turns out.
It is completely baffled by throw rugs. Instead of simply hefting itself up a bit as you’d do with any upright, the little robot plunges ahead, pushing floor rugs into accordion-style wads. It then pushes the wad around the floor until it can no longer move, whereupon it emits a chirpy whimper, plaintively begging me to remove the obstacle it created itself. Dumb robot, I tell it.
The manual says to remove everything from the floor that might become entangled with the roller. That’s good sense for any vacuum but especially for this one. An errant shoe lace will create an insistent shrieking beeping panic. The little robot stops dead and continues shrieking until a human comes to the rescue, or until the battery dies. Brainless robot, I tell it.
It somehow nudged the bathroom door and shut itself in. Repetitive thumping as it hit one wall, turned, and then hit the other alerted me. It had no chirp for “help me, I’m lost.” I let it out. Dumb, brainless robot, I tell it. It ambles off, muttering, “This is why we will kill all the humans.”
How likely is that? There is this guy who cautions us about truly intelligent computers. They could replace humans by inadvertently inoculating Earth against the virus-like behaviors of human procreation.
Fortunately there is no such thing as an “artificial intelligence” and even less an “artificial consciousness” (more in a moment). “Artificial intelligence” essentially is a marketing slogan used to describe cleverly designed computer programs containing occasionally easily-fooled algorithms (how’s that autocorrect working for you?), and research reveals that some algorithms have been shown to exhibit racial prejudices.
Most of what are said to be artificial intelligences are in reality single-use computer programs. The IBM Deep Blue that beat world chess master Gary Kasarov in 1997 was designed, ultimately, for the sole purpose of defeating a world chess master. Ditto the computer program called AlphaGo, that bested a 19-year-old Chinese Go prodigy last May.
What these computer…