Today, micro-decisions and information can be anywhere — but the daily work that keeps humanity humming cannot be. This is old-hat logic from futurist Alvin Toffler: Muscle power is not fungible. But as he forgot to say, muscle power is necessary.
The simple fact is that the tasks we do every day will not be replaced by AI for a long time. It’s almost 2018, 50 years after “2001” the movie was released, and we still have nothing even close to HAL — who, by the way, was dealing with a simple problem space.
The dystopian world predicted by the pundits won’t happen until an AI can fix a leaking toilet at 524 73rd St. Brooklyn, N.Y., right now. Until then, a plumber will make more money than a strategy consultant, and probably should. If you are a normal person, this will seem obvious to you. But why, then, other than tabloid motivations, does AI get so much press?
Today, most press releases regarding AI are little more than shameful promotion of computing that can do a human-like trick, usually badly. For algorithmic tasks like chess, fine, the computers win. For information retrieval combined with voice recognition, like on “Jeopardy,” same thing, but Alexis can’t even credibly tell you where to go to dinner. Even IBM, vaunted AI promoter, backed off using the term AI in favor of “cognitive computing.” That wasn’t a pivot so much as it was duck and cover.
Cold Dog Soup and Rainbow Pie
The fears of rational experts are more about the harm that a machine might do with the elements it can control. Last month a New York Timesop ed by Oren Etzioni, an eminent AI researcher, proposed laws as upgrades to Asimov’s Robots' Rules of Order.
The article pointed out the practical ways that AI might harm us. We should be cognizant of those, and respond in a measured way.
First, we should make it illegal for a machine to do anything that a human can’t do. Then, we should make a human accountable for any point of control, from the simple things like home automation, to the complicated ones like control of a nuclear power plant. This will, at least, deter pathetic defenses such as airlines blaming pre-hurricane price gouging on “the algorithm.”
Blaming anything on an algorithm is the worst kind of sniveling cowardice, but reveals an inconvenient truth: If companies understand the algorithm well enough to know what it could do, shame on them. If they don’t, shame on them.
But still, Elan and the soothsayers (great band name!) are doing society a disservice by promulgating fear without proposing solutions. As I write this, Ray Kurzweil was quoted today saying he’s not worried about AI taking jobs. This from the guy who brought us the Singularity. Kurzweil v Musk. LMK if you want my tickets.
AI will create conveniences, it will help us solve problems, and it may even aid in decisions that will stop global warming or fix my toilet. But it won’t stop global warming, and it won’t fix my toilet. Only humans can do that.
Forgive me for harping on the obvious, but to iconize or demonize computers, or aliens, Russians, immigrants or any other hobgoblin is ducking accountability. If we don’t own the massive issues we face as a society, if we don’t look them in the eye, we have only ourselves to blame. Whatever dystopian future you may imagine, the people there will look back and blame us, not computers.
The paranoid belief that AI will destroy us is just as neurotic as the false hope that it will save us. These are not ghosts, they are machines — and, at least for now, we make the decisions.