Posted: Jon Peddie 11.03.17
As the wedding between GPUs and AI move from the honeymoon to the serious setting up of a household and becoming a member of society, it will have to find a new name for itself. The term artificial intelligence, and AI, strikes fear into the psyche, heart, and groin of most of the people beyond its developers. Morgan, Ava, David, The Terminator, and the replicates to name a few have given AI a bad name; fun in the movies, but scary as hell in your home. Always listening Alexa, and Siri, and the new wave of always watching home assistants is promising even more intrusion into your life in the name of assisting you. Following the science-fiction model, you arrive home after work all tired, your home assistant takes notice , selects the best music for you, prepares an appropriate drink, adjusts the lights and temperature, and suggests in a soothing voice you sit down, take off your shoes, (the assistants don’t have olfactory sensing yet) and asks if you would like to watch TV, subsequently selecting the appropriate channel for you.
What’s wrong with that? Nothing. Few, if any of us, have a human butler that might do all those things, so this would be a definite improvement in the quality of life. But, what if you came home late with a date, does your AI butler offer you a sexy movie and dispense a contraceptive because it has read your heart rate, and seen your flushed face? How would your date react to the presumptuousness of the AI butler?
However, between here and there we have to find another way to describe these devices. Also, the term intelligence is not only an overstatement, it is not accurate. Intelligence is defined as the ability to apply knowledge to manipulate one's environment or to think abstractly as measured by objective criteria. Our AI toys today can apply knowledge about a very limited number of things, simple command words, and images. The AI that knows how to get the current temperature in a foreign city may not know where in that city the temperature sensor resides. If not, it can’t tell whether it’s in a windy canyon between skyscrapers or at the airport. I don’t think there’s much need for discussion about the ability of the current, or next generation of AI devices to think abstractly; not many people can even do that successfully.
Nonetheless, we have been raised on a steady diet of scary as hell AI creatures through the movies starting with Maria's robot double in Metropolis in 1927, and by automatons at public events before that.
In most, almost all of those movies, the AI was either initially or in the end evil, and either threatening or merciless. Ah, and there is another judgment of AI, can an AI make a moral and/or kind decision?
The instances of AI characters have also inundated science-fiction literature since the 16th century with Golem, and the Metal Men in 1899 (Gustave Le Rouge’s La Conspiration des Milliardaires—The Billionaires' Conspiracy). It wasn’t until 1920 that Karel Čapek coined the term “robot.” Television picked up on AI and robots early in its life and the first shows started appearing in the early 1950s, with the lovable Robby appearing in 1958 following his introduction in the great movie, Forbidden Planet in 1956. For the most part, Asimov’s positronic brained robots were benign, following stringent rules of behavior, but that was a passive-aggressive way of pointing out that these were dangerous things that needed to be strictly regulated and managed, something that has been called for by seriously smart leaders in business and academia today, and as soon as it is discussed on Fox TV, our political leaders will get on the bandwagon too.
Autonomous vehicles, land, sea, and air, employ AI but use the term less often, if at all. Those are big, expensive, and scary machines, and the association with an errant AI, or an AI that’s in a foul mood, is too scary to think about when you’re in the back seat of one whizzing down the autobahn at 200kmph.
My kids used to say I never asked a question I didn’t have the answer for, but this time I don’t. I don’t really have a suggestion as to what we should call these independently functioning semi-autonomous rapid decision-making teachable devices; I’m sure there’s an acronym in there somewhere for the marketing-inclined.
Of all the AI characters, real (Alexa, Siri, etc.) and imaginary (i.e., films and books), which one do you think is the scariest, and which the most benign?