第140页 | What to Think About Machines That Think | 阅读 ‧ 电子书库

WHAT WILL AIs THINK ABOUT US?

GREGORY PAUL

Independent researcher; author, The Princeton Field Guide to Dinosaurs

The following discussion presumes that conscious minds function in accord with the laws of physics and can operate on substrates other than the neurological meat between our ears, that conscious artificial devices can therefore be constructed and will become even more self-aware and intelligent than humans, and that the minds operating in human brains will then become correspondingly obsolete and unable to compete with the new mental überpowers.

This particular primate-level thinking biomachine thinks the development of artificial superminds is a good idea. Although our species has its positives, Homo sapiens is obviously a severely limited, badly designed (by bioevolution) system that’s doing grave damage to the wee planet it inhabits, even as the planet does grave damage in return—e.g., diseases have slaughtered about half the some 100 billion kids born so far. Attempts to preserve humans indefinitely into the future much as they currently are is a conservation project that flies in the face of evolutionary processes, in which species come and go in a continual turnover. There’s no a priori reason to presume that we’re so special that we deserve exceptional protection, particularly if our successors are capable of self-aware, conscious thought.

But to be blunt, what we think about these issues probably doesn’t matter all that much; humanity as a whole is not really in charge of the situation. Once upon a time—the year 1901, when my grandmother was born—building flying machines was so hard that no one had yet done it. Now the necessary technology is so readily available that you can build an airplane in your garage. Once upon a time—shortly before I was born—we didn’t understand the structure of DNA. Now grade school kids do DNA experiments. Currently, the technologies needed to generate nonbiological conscious minds aren’t on hand. Eventually, commonly available information-processing technology will probably become so sophisticated that making thinking machines will not be all that hard to do. And lots of people will want to create and/or become cyberminds, no matter what others might think and despite what laws and regulations governments may pass in futile efforts to prevent their onset.

In the end, all the contemporary chit-chat about the cyber-revolution often called the Singularity is so much venting and opinionating, not all that different from the pretty useless discussions back in the 1800s about the feasibility, advisability, and ultimate meaning of the advent of powered flying machines. What we say now doesn’t count for much, because if the technology never works, then superminds will never be either a problem or a benefit; and if the technology does work, then one way or another the new thinking machines will be devised and will take over the planet whether we like it or not.

If so, the important question won’t be what we think about thinking machines, it will be, What do they think about old-fashioned human minds? One item there’s no need to fear is the enslavement of hapless humans by their cybersuperiors. People are too inept and inefficient for smart robots to bother exploiting them. Even now, corporations are trying to minimize the labor they have to pull out of big-brained primates. The way for human minds to avoid becoming obsolete is to join in the cybercivilization by uploading out of growth-limited biobrains into rapidly improving cyberbrains. That could be for the best. If high-level intelligence can get out of the billions of human bodies weighing down the planetary ecosystem, the biosphere may well return to its prehuman vitality.