July 29, 2009
Smart machines: What's the worst that could happen?
MacGregor Campbell in New Scientist:
An invasion led by artificially intelligent machines. Conscious computers. A smartphone virus so smart that it can start mimicking you. You might think that such scenarios are laughably futuristic, but some of the world's leading artificial intelligence (AI) researchers are concerned enough about the potential impact of advances in AI that they have been discussing the risks over the past year. Now they have revealed their conclusions.
Until now, research in artificial intelligence has been mainly occupied by myriad basic challenges that have turned out to be very complex, such as teaching machines to distinguish between everyday objects. Human-level artificial intelligence or self-evolving machines were seen as long-term, abstract goals not yet ready for serious consideration.
Now, for the first time, a panel of 25 AI scientists, roboticists, and ethical and legal scholars has been convened to address these issues, under the auspices of the Association for the Advancement of Artificial Intelligence (AAAI) in Menlo Park, California. It looked at the feasibility and ramifications of seemingly far-fetched ideas, such as the possibility of the internet becoming self-aware.
More here.
Posted by Abbas Raza at 01:33 PM | Permalink
Comments
So the Cylons will soon be among us!
Seriously, once we have devised machines for mindless lack of common sense and violent irrationality, they will have become self aware (i.e. like us)!
Posted by: Bill | Jul 29, 2009 4:41:22 PM
I had a thought the other day (hey, it happens).
If we are ever visited by extraterrestrials, the odds that they will actually be machine intelligences are nearly perfect.
Any civilization advanced enough to field a probe to a distant star will surely have extremely advanced AI technology, and no one else could survive the trip.
Iain M. Banks' "Culture" novels' AI overlords who can't seem to recall the organic intelligences that created them are a thought provoking 'meme' that deserves fertile soil. Pay special attention to the moral code as applied, not as promised. Very subtle, and creepy. Nice folks, though.
Posted by: Carlos | Jul 29, 2009 11:17:36 PM
I await the war between AI and genetically modifies superhumans.
Posted by: Robin | Jul 30, 2009 1:35:14 PM
Asinine Innovation
I cannot see anything in this article that merits its inclusion on this web-site. It says nothing new and does so in a smarmy school-boy style.
1. The term AI is acceptable to me; it is something computer wonks do and make a lot of money doing. The term Artificial Intelligence (capitalized) is jargon, something I am forced to live with and cannot fight. The term artificial intelligence is silly enough, I just won’t put up with ‘artificially intelligent machine’, or ‘conscious computer’, or ‘human-level artificial intelligence’, or ‘roboticist’, or …
2. So, the world’s leading AI researchers have been discussing the risks over the past year and have now revealed their conclusions. Oh thank God for that. What are they? Damned if I learnt anything from reading this article that I hadn’t already surmised.
3. Roombas, Scoobas, a lifeguard aid, a route selector – these are AI devices? These machines are implementations of ideas by intelligent humans who believe they have considered all possible situations in a particular application, and have programmed miniature computers with instructions on how to deal with each situation and control a mechanical or electronic device appropriately. What is all-fired AI about this?
4. The panel of experts warns of machines of the near future that make and execute decisions on their own, albeit within a narrow range of expertise. The actual words in the article are ‘the day is not far off’. Has this guy ever tried making a complaint about his phone service to the phone company over the phone? I need a panel of 25 highly paid experts thinking their heads off for a whole year to tell me this drivel? These are the folks who made feasibility projections some where between 20 and 1000 years! A savant idiot could have prognosticated as precisely.
5. In this post-Rainman world is there a need to put idiot savant in quotes? A smarmy school-boy would! As for the plural, if you wish to impress someone with your facility with French, you would print idiots savants in italics. Webster’s anything-goes acceptance notwithstanding, idiot savants suggests a bunch of wise men who are mentally challenged. Anyone with human-level intelligence would know that idiots savant is correct. And when did we stop being proud Anglo-Saxophones and start pluralizing adjectives to qualify plural nouns?
6. But try explaining syntactical simplicities to the ‘we could – they could’ idiot savant who confesses that all the smarts of CMU AI Labs is not any better than the mobs.
7. Why would anyone characterize a chain reaction as a singularity? More euphemisms for the future perhaps, like China Syndrome.
8. What is the digital behavior of humans? The way we scratch our butt-cracks?
9. Finally we have : ‘Given such possibilities, "what's the responsibility of an AI researcher?" says Bart Selman of Cornell, co-chair of the panel. "We're starting to think about it." At least for now we can rest easy on one score. The panel concluded that the internet is not about to become self-aware.’ Thanks guys. Perhaps if we double your per-diem you’ll come back next year and tell us how you spent your summer vacation.
Really Abbas, you should be more discriminating about what you pluck off the Internet to post on 3QD. Don’t you have a God-given bullshit meter? Here’s a simple rule of thumb that invariably works for me: If the author has two last names then skip the article. (How's that for rule-based decision making?) It is surely a good enough reason to trash this one.
Posted by: narayan | Jul 31, 2009 5:17:36 AM
Post a comment