132
8 Biomorphic Technologies
puter games might be precursors of this Brave New World; with their way of life and
the education they get they will need AI to care for them – but why should AI care
about maintaining obsolete electrochemical appliances, much less providing them
with inexpensive vehicles with nanoengineered microwings (which, as Kurzweil implies, should be able to carry people not yet miniaturized to microbes)? Why would
it care for hungry uneducated masses? Would it maintain zoos housing creatures
stubbornly clinging to their antiquated habits – Buddhist monks, Amish, orthodox Jews, recluses, remnants of hunter-gatherer tribes? Homo Deus by Yuval Noah
Harari (2016), predicting the same kind of future, recounts a grim dystopia, whether
intended or not.
The acolytes of exponential growth, while ridiculing linear forecasters, commit
a similar error: they expect the exponent to remain constant, but it cannot. When
any process appears to be rushing to the skies, something must happen on the way.
In the best case, the exponential growth saturates, and something new and unpredictable emerges, and perhaps starts growing exponentially itself. In the worst case,
the system explodes. History tells us that many prophecies have never materialized;
the future is hard to see, as a simple-minded song reminds us, and still harder for
professional prophets with their ingrained opinions and prejudices.
The image of human-like robots ubiquitous in science fiction is naive. The superintellect of Solaris (Lem, 1961), mentioned in Sect. 7.4, may reside not in the structureless ocean but in a post-Singularity machine hidden underneath, either evolved
from or created by its extinct biological precursors. On our planet and in our time,
worldwide AI, at least in an embryonic form, already exists, hidden in Google’s
computer network and its rivals; it collects information about our movements and
actions, and already surreptitiously governs the world. Self-driving cars, domestic
robotic appliances, air navigation systems, you name it, are extensions of a remote
AI. Their in-built processors, like the ant’s brain, are too weak for their range of operation: for example, a car’s computer couldn’t hold the city map and traffic density.
Like worker ants, they are parts of a remotely interacting assemblage, but, unlike
ants, they do not self-organize, but are directed by the mighty omniscient remote
processor. Only protected networks disconnected from internet for fear of hacking
are out of reach.
AI still needs human help for maintenance, but takes over much of the programming. It also directs scientific research and development, serving its needs through a
scientific communication network. A run-of-the-mill scientist is as minute a part of
this network as an ant in its anthill. The paradigm-changing discoveries of the last
century are receding into the past. There is a hope, at least, that scientists, like ants,
self-organize – but do they? Or only in their private lives? The paths adopted by
their foraging are laid out by funding agencies, and by university and government
bureaucracies which, lacking independent judgement, are directed by the same remote processors collecting mountains of trivial statistics.
Biomorphic technology may eventually bring us brain-enhancing implants and
rejuvenation treatment for the rich, while the masses failing to reach the Homo
Deus status will, in the best (for AI) case, be drugged by electronic entertainment
(as many already are), or in the worst case, rebel, bringing down the whole house
Précédent

- 137/151

Suivant