Thursday 6 June 2019

Consciousness-Level Computers: Issues

The People of The Wind., V.

Daniel Holm argues that the Terrans might get a robotic spaceship to:

"'...land like a peaceful merchantman. Consciousness-level computers aren't used much any more, when little new exploration's going on, but they could be built, including a suicide imperative. That explosion would be inside a city's force shields; it'd take out the generators, leaving what was left of the city defenseless; fallout from a dirty warhead would poison the whole hinterland.'" (p. 491)

Two Issues
AI
Rule-governed manipulation of symbols is neither knowledge of their meanings nor consciousness of anything else. Therefore:

computational processes, whether mechanical or electronic, merely simulate but not duplicate intelligence;

computational processes that are merely more "...sophisticated..." (p. 491) will not thereby rise from an unconscious to a conscious level.

However, an artifact that duplicated brain processes would necessarily be conscious and such an artifact might also perform computational processes.

Morality
Morally, to build a conscious and intelligent artifact with a suicide imperative would be to commit murder.

1 comment:

Sean M. Brooks said...

Kaor, Paul!

And none of our computer systems (hardware/software) has duplicated the mind like that. And I am skeptical that is even possible. Moreover, the Imperials did not use a ship with an AI programmed to have a suicide imperative. Which argues for the Terrans deciding there were some things they could not decently do even in war time. So Daniel Holm was wrong.

Sean