Tuesday 7 July 2020

The Beginning Of Conscious AI

Poul Anderson, Genesis (New York, 2001), PART ONE, III, 2.

Christian Brannock is remotely neurally linked to the robot, Gimmick, which is exploring the Mercurian surface. Brannock's colleague, Willem Schuyten, reflects that:

"He had often heard such highly developed computers and neural nets, with their programs and databases, called 'brains.' People who worked with one, like Christian - although seldom as intimately as he did - were apt to give it a name and speak of its personal quirks, as other people might speak of a ship or a tool that had served them a long time." (pp. 15-16)

This is where a crucial transition occurs. Conscious beings use tools, including computers, but the tools are not conscious. Neurons in brains generate consciousness. If a "neural net" duplicates the functions of neurons, then it also generates consciousness. Thus, conscious artificial intelligence emerges.

The name, "Gimmick," enables us to make a comparison with the soft sf writer, CS Lewis, who writes a convincing account of a Lunar landing but displays his lack of scientific knowledge and understanding by dismissively labeling scientific instruments as "gimmicks":

"The landing was not without terror, but there were so many gimmicks to look after, so much skill to be exercised, that it did not amount to very much."
-CS Lewis, "Forms of Things Unknown" IN Lewis, The Dark Tower And Other Stories (London, 1983), pp. 124-132 AT p. 128.

In imagination, we venture to the Moon with Lewis and to Mercury with Anderson.

1 comment:

Sean M. Brooks said...

Kaor, Paul!

And I still have very strong doubts that it will even be possible to create or build an artificial neural net duplicating the functions of human neurons, hence an "AI." I believe Anderson was also skeptical but was quite willing to speculate about such things, as we know.

Ad astra! Sean