|
Haha, this is a joke.
I'm no expert on robotics, but I presume that the level of technology isn't quite near enough to be competitive with the dexterous primates we are. We are nature's perfect manual labourers and precision-coordination instruments. This, then, takes care of the flipping burgers bit (or not). (Personally, though, I wouldn't be too worried about McBot taking over my burger-flipping job, because it sucks. This is not what you're all worried about, I presume.)
However, as far as the AI camp is concerned, I do know a thing or two, and I can tell you that automated McDonalds servers are NOT going to be showing up for a long time (I'm not talking about automated ATM kiosks, which are not hard to do -- let's look at the bigger picture here). It's quite simply just too complicated to implement. Even something as simple as ordering a damn Big Mac, believe me. Just to provide a simple example -- say someone is ordering something from an automated McDonalds station -- say it's a mother with a small child. The conversation goes as follows:
McBot: "How may I help you."
Mom: "I'd like to order a Big Mac and a Kid's Meal."
McBot: "What would you like to drink?"
Mom: "Hmm, tough question, I think I'll go with Coke. But I don't know about the kid. Let's see... *ouch*! Honey, stop pulling! Yes, we're going to eat soon, just please be patient! Ahh just scrap the soft drink for her, we'll just take juice."
McBot: "......"kaboom!
Do you realize how hard it is to process this conversation? Assuming that a machine can perfectly hear what you are saying and render it into speech, how will it know what you mean by "I'll have coke but I don't know about the kid"? How can it know that by that you mean that you are getting the kid's meal for "the kid"? How does it know it's YOUR kid? How does it know to wait for you to find out what your kid wants, because how does it know that it matters what your kid wants? It expects a straight answer. And how does it know how to distinguish between relevant information and garbage, like that little "pulling-on-mommy's-clothes" interlude? How about the "scrap the soft drink we'll have juice instead"?
A classic problem like this in the AI field involves an "AI parser" reading a preschooler story about a birthday party. The story goes (roughly) like this:
John was having a birthday party. All of his friends were invited. Sally was unsure of what to get him. She ended up going with her mommy and buying John a kite. But when they got to the party, and John opened up his present from Sally, he was sad because he already had a kite, and had to return the new one from Sally.
Any kind of computer that would have to read this would have to be aware of a staggering amount of information:
- what birthday parties are
- why it is the custom to bring presents to whoever's birthday it is as opposed to the host giving out presents to the guests, or not having presents at all
- why people invite friends to birthday parties as opposed to total strangers or even enemies
- what the emotional significance of birthday parties are (goes in line with #1, above)
- what emotions are (cut your teeth on that one!)
- why birthday parties are attended at parties and not at business meetings or charnel grounds
- why John was sad when he found out he got a kite he already had (goes in line with what emotions are)
- why it is the custom to return presents or exchange them if you already possess a copy of it, as opposed to keeping it and having two copies
- why Sally would go buy a present with her mommy as opposed to going alone or going with a stranger or her cat
Some of these may sound funny or stupid, but it just comes to show how much BUILT-IN knowledge (which we all take for granted!) must be assumed to understand this simple, one-paragraph mini-story that ANY FOUR YEAR OLD COULD UNDERSTAND.
The point of this long ramble is that the image of robots taking over humanity's work is FARFETCHED. Machines are good at doing tasks that are SIMPLE and REPETITIVE, like adding numbers quickly or building cars. They CANNOT perform tasks like communicating with humans, understanding language, conversing with humans, working in any kind of service area like fast food or anything of the sort. Human language, communication, and thought are just so immensely complicated that we are decades, if not centuries, away from creating anything that can understand things at an appreciable level, much less at a level that is COMMERCIALLY VIABLE enough to totally REPLACE human workers. And for THAT kind of a level, you need EXPERT systems that can compete if not EXCEED the performance of humans.
So don't worry, the Matrix isn't here just yet...
Last edited by Alccode on Aug-13-2003 at 06:00
|