5.11.2005

 

Conscious airplanes

The White House and Capitol in Washington DC were evacuated today because a small Cessna 150 airplane strayed into restricted airspace:In this case it appears to have been a complete accident -- the pilot was a student.

Within 10 years or so, airplanes will start being equipped with a rudimentary form of consciousness that makes this sort of accident, as well as intentional attacks, impossible. The book Manna, chapter 3, describes the progression.

The system will work like this. Airplanes can already detect their exact location using GPS systems. These GPS systems will be married to very detailed digital maps of the ground and the airspace over the ground. The maps will tell the airplane where every single building and structure is on the ground, and mark all areas of restricted airspace. So if a pilot steers a plane toward a building or a restricted area, a "conscious" plane will refuse to fly there. It will be, quite literally, impossible for a conscious plane to fly into a building -- the plane will "know" that flying into a building is "wrong." If all the engines fail, a conscious plane will know what is on the ground in the vicinity and do its best to crash into an unpopulated area.

As this happens, of course, planes are able to fly themselves. Pilots go extinct, and airplanes become completely automatic. Human pilots are eliminated from the cockpit because they cannot be trusted. See Manna and Robotic Nation for details.

Comments:
Don't you mean Conscience? (Sense of "right" and "wrong"?)

Consciousness would be if the plane were to be having an experience, not just computing. Chalmers territory, and the like.

-- Lion
 
Perhaps conscience can only come from consciousness. The ability to resolve your own self as distinct from your surroundings may be the ultimate arbiter of what you find right and wrong.
 
Hm... I believe my computer sees itself as distinct from it's surroundings.

For example, if it sees a message addressed to 127.0.0.1, the computers way of thinking "I," it will just give itself its own message, and not try to send it out. If it sees a message addressed to a computer with a different IP address than it's own, then it will send the message over ethernet or the wireless.

So, it seems that my computer meets the condition of seeing itself as distinct from it's surroundings.

But I would not say that it is conscious, nor would I say that it has a conscience.

I would personally reserve the term "conscience" for an AI that was able to converse, that would try to persuade someone that an act was right or wrong (of it's own volition,) and that could be persuaded that an act was right or wrong (which it must be willing to hear out, in some situations.)

Consciousness, on the other hand, seems to me altogether different.

I would personally trust that an artificial intelligence were conscious, if it met the following conditions: (1) It was not programmed, directly or indirectly, to say that it was conscious. (2) It nonetheless insisted that it was conscious. (3) It had the same difficulty telling what consciousness was, that a human has. OR: The artificial intelligence was structured very similarly to another that met those three conditions.

I can imagine an AI that has conscience without consciousness, and I can image an AI that is conscious, but does not exibit conscience.

-- Lion
 
Post a Comment

<< Home
Archives © Copyright 2005 by Marshall Brain
Atom RSS

This page is powered by Blogger. Isn't yours?