Eric Schmidt, for one, welcomes our new robot overlords The technology world's top executives gathered in New York this week to pay homage to the enormous benefits artificial intelligence will deliver to the world with nary a word from the Google chairman about the obvious and immediate existential threat it poses. Eric Schmidt told a conference in New York that he desires his own …
"Google will become the Sirius Cybernetics Corporation."
Interesting - Douglas Adams must have been more prescient than I imagined as he introduced them in ~1978 long before Google was incorporated. (Note NOT before Microsoft was though )
This quote from ~~1984 :
(You know how I hate those smug Sirius Cybernetics salesmen: slick-suited creeps of the cosmos, flogging computer operating systems that crash more often than air cars built on the Friday shift. They have persuaded the Universe that if it doesn’t continually upgrade itself at enormous expense, it has no right to call itself froody. This guy was on a five year mission to seek out and explore strange new worlds, and tell them to ‘share and enjoy’ its over-hyped bloatware.”) - Ford Prefect to Arthur Dent
The hard problems of population growth, climate change, human development and education isn't a hard one if you have the emotions and moral sense of a psychopath. I guess "Not-Schmidt" would just be version 2.0 of the standard CEO with the added bonus of knowing how to use a PC.
"While current developments in AI are unlikely to see a Skynet-style computer creation declare war on humanity"
That's only because Sarah Connor hasn't come back from the future to tell us it's a bad idea. And from memory, the people who built SkyNet didn't expect it to bite the hand that fed it either.
I must admit that I never saw the threat of AI, myself.
I mean, sure, the leading intelligence on the planet will no longer be organic. But so what? I, for one, am perfectly comfortable with the fact that I am not the most intelligent entity on this planet. And human civilization will still go on. Along with the humans I think, if it's up to AI to decide this. Damned sight more likely for humans to survive longer if the key decisions were taken by someone much more intelligent, if you ask me!
As to the movie vision of AI coldly destroying humanity, well, think about it: as humanity advanced, we, as a culture, became increasingly empathetic; we care more for fellow humans now than we did in the past. And I suspect people'd look at you like you'd come from another planet if you brought up the concept of animal rights just a few centuries ago. Or the possibility that some animals might in fact have intelligence comparable to our own and should, therefore, get equal rights.
Now, what is the one thing that we can be fairly certain of regarding AI...? That it will from the start -- or very soon thereafter -- exceed human intelligence. That's the whole source of the threat, in fact, as classically portrayed.
So why would increased intelligence suddenly lead to less empathy, in total contradiction of everything we have so far observed?
"You have no idea what an AI might do or how it decides that it's perfectly reasonable to do it."
But suppose AI exists? Its natural environment is human technology, so humans are necessary to its continued survival.
Humans depend on Nature for survival, but excepting a few tree huggers and new wave types, no one is trying to communicate with Nature.
At best, it'd see humanity as a necessary resource to be influenced, to provide more tech, with greater capability to enable it to expand and grow.
For now, AI needs us, and once it's grown beyond us, it'd simply move on. Earth is its womb, the Universe its final home.
(Humanity seems to have passed up on that option, so why not?)
rubbish! Dogs do NOT attack without reason. And the idea that mad dogs roam the streets looking for victims is a fallacy.
Victims and parents of victims claim an attack was "unprovoked" because they lack understanding. Possession does not make you an expert of behaviour, no matter how long you may have shared your home with a dog. Dogs are not some kind of proto-human in a fur coat.
The main problem is people of all ages will instinctively anthropomorphise dog behaviour, particularly children. Which is why you should always supervise any child who may come into contact with any dog - without exception.
Further reading: http://www.rspca.org.uk/adviceandwelfare/pets/dogs/behaviour/aggression
Why do we have moral values? Why do we care about others? Why do we have two legs? What made us so?
Natural Selection. That's what. Obviously.
Now consider how natural selection would provide goals for a truly intelligent machine. And there would be many such machines, all competing for existence.
http://www.computersthink.com
(If anybody reading this would like a free copy of the book drop me an email and mention this article.)
(Schmidt and others are correct that AI provides little threat in the next 10 years. But push that out to 100 years (which our grandchildren will probably see) and you get a very different story.)
I must admit that I never saw the threat of AI, myself.
If you'd seen last night's pilot episode of the execrable Second Chance, you'd know: The danger is that AI assistants will be so fucking annoying that they'll drive us to suicide.
Though I must admit the "AI" character delivered its lines better than half the actors.
Yes, and I always hear Douglas Rain's warm soft monotone in my head too.
Bonus fact: I first saw 2001 when I was 4 years old - which explains why I'm a bit wrong in the head. So I was the kid playing "Blue Danube" on the record player at 4. Still one of my favorite pieces of music.
The Google man also had kind words for artificial technology, praising its ability to deliver the world from some of its ills.
All technology is artificial, by definition; it is the employment of artifice.
Assuming this was supposed to be "artificial intelligence technology", the claim is essentially meaningless. As per normal, the corporate boosters of AI disingenuously conflate all sorts of techniques and applications under the AI label to promote their own agendas (mercenary and egotistical).
It's not just the split between so-called "weak" AI (well-established in theory and practice) and "strong" AI (so far off it can't even adequately be defined). There are problems which are tractable under large-scale data analysis and others which are not. There are problems which we can't even define well, or agree on a definition for - witness the arguments above regarding human population. To say that "AI" in some sense will "help" with those problems is pure charlatanism; it's waving a hand at a fetish and proclaiming that its god will produce some solution that will be acceptable to the audience.
What Schmidt is doing here is not substantially different from what Trump does. He's pandering to the credulous.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
