Inevitable Positive Outcome with AI?

For those who believe that human-level AI isn’t far off and that a rosy scenario isn’t inevitable, 2009 is a somewhat sad and depressing time. Popular opinion is that AI won’t be here for centuries, but that isn’t a huge problem or issue. (In fact, it makes things easier by limiting the number of people involved in AI research, thus allowing me and my confederates to keep a closer eye on them.)

...

Complete entry

Posted by Patrick  on  01/09  at  04:55 PM

It's interesting you bring up day-trading as an example. The difficulty with all automated trading logics, even those backed by sophisticated and adaptive algorithms, is that human irrationally en masse is too chaotic and non-linear to adapt to without some large-scale modeling of the trading population. Therefore, the real impediment to that is data (the stream of where volume is going) or capital (the ability to move the market enough for systematic and adapting manipulation).

At this point, maybe a day-trading AI is the best ticket. Maybe the AI can apportion resources to be deployed by humans, or roll up the whole global markets and give it all out fresh.

Posted by PredictionBoy  on  01/10  at  05:04 AM

I agree with much of this article, especially regarding the difficulty of "fully" integrating AI with the human mind. (Whatever that means)

However, it seems to imply that if this can't be achieved, then we may destined to standalone, "monomaniacal" AI, simply because it isn't fused with human brains. This is giving in to science fiction ideas that make little sense in the real world.

In fact, a case could easily be made that AI that is fully fused with the human mind would actually be more dangerous than stand-alone advanced AI, because a fully-fused AI/human brain construct would presumably have a dominant or at least strong component of irrationality that currently drives human brains - because our evolutionarily-derived brains would be part of this construct, and evolution rewards irrational, dominant, often sexually-oriented behaviors. Evolution rewards this basic behavior pattern not just with humans, but most creatures on earth.

But AI will be a designed technology, very different in its inherent nature. In some ways immensely more powerful, in other ways motivationless, soulless, arid - but that's not a bad thing, because these will be tools, like all technologies, not some deliberately created "replacement species" for humans, that is a ludicrous concept.

Stand-alone AI would almost certainly be driven by a form of rationality far purer than anything we are familiar with in our minds. The idea of a "monomaniacal day trader" that "decides" in a fit of irrational dominant impulse to take control of the entire stock exchange is a deeply improbable sci-fi fantasy.

But again, most of the article describing the difficulties of fully integrating AI and human brains, and criticizing the breezy way that Kurzweil and others predict this will come to pass, is good content.

Posted by Sam Kenyon  on  01/13  at  11:53 PM

"1) Most new technologies are created as stand-alone objects."

Stand alone from what? Of course from direct physical interaction with the human brain (so far). But not from human brains indirectly, and not from other objects. All tools and entertainment techs are extensions of and/or interact with the human body, either physically and/or with information processing (the brain). Technologies range from completely human controlled to completely autonomous. Badly-designed user interfaces, catastrophic bugs and malicious programming cause technology to be "naughty." So, I don't see how the difficulties of neural interfaces has anything to do with the naughtiness of intelligent machines.

---

Add your comment here:

Name:

Email:

Location:

Remember my personal information

Notify me of follow-up comments?

Please enter the word you see in the image below: