The current topic, in specialized media and digital sites but also in
mainstream ones, concerns the so-called "artificial intelligence" (AI),even if this definition seems to cover a large number of phenomena, from
the washing machine at home to self-driving cars, without non-specialist
readers understanding much about it, except those who grew up reading
Asimov and Clarke. I don't know if this is the case for the nerds who
are behind the latest acceleration of research on the topic and its
practical applications, but they are certainly well-trained in physical
and mathematical sciences, much less in humanities, linguistics and
semiotics, for example. This would seem to indicate that the path that
perhaps will lead to the "thinking machine" begins in very distant
times, from the creation of tools for cultivating the land or devices
for waging war until arriving, in the Western case, at automobiles or
typewriters; a path based on what Horkheimer called "instrumental
reason", a peculiar logic of the relationship between means and ends,
subservient to particular interests, characteristic of our capitalist
society, born from the industrial revolution. The coherent and logical
conclusion of Western modernity, based on the individual as a single
subject who thinks and who objectifies himself outside of himself in his
own instrumentation, in a game of mirrors that gives him meaning: AI,
then, seems thinkable only within modernity, understood as Western.
The problem with this conclusion concerns the fact that there are traces
of "thinking automatons" also in societies of the distant past and not
only Western ones. Certainly we must distinguish between the idea of the
automaton contained in fantastic tales or mythologies, where the
anthropomorphic object is animated thanks to spells or magical rituals,
and that recorded in annals or historical accounts, where instead it is
the mechanical technique that justifies their behavior. These data
demonstrate the presence of a technical development capable of producing
statues with movements, which could also dance and sing, as in the case
of the Chinese mechanical automaton built in the 10th century BC by Yan
Shi, a craftsman of the Western Zhou dynasty. But it is in classical
Greece that we encounter abundant and ancient references to specialized
automatons, the construction of which is attributed to Hephaestus, such
as the "golden dancers" and self-propelled tables, even if his most
famous automaton was Talos, a bronze warrior who defended Crete and who
was defeated by Medea who accompanied Jason, ensnaring him with her
magical arts and unscrewing the screw that was in one of his feet, from
which the "liquid" that provided him with the energy to move flowed. If
we consider the complex gears of the "Antikythera machine" (a
planetarium) found on the seabed of the island from which it takes its
name, it does not seem so strange to us that between the fourth and
first centuries BC there existed in the Mediterranean a mechanical
technology so sophisticated as to build "programmed" automatons to
perform complex actions.
Considering these data, it is possible to solve the problem I posed on
the relationship between modernity and AI, of which Greek automatons can
be considered the ancestors: I think, and I am not the only one, that
during the four centuries before the beginning of our era, in Greece
social and cultural transformations took place that were about to take
the forms of what we call "modernity" today (which means that the same
fluctuations may have manifested themselves in other regions of the
world); progressive slippage from the interpretation of the world, from
religious to a more materialistic one, emergence of the individual
subject, representative political systems, etc. In this context, it is
not unusual that in the post-Platonic philosophical world, which already
had "modern" interests in physics and biology, one could discuss
automatons or mechanical animals such as the steam-powered flying dove
mentioned by Homer, from which Aristotle derived a reflection on the
impact of the development of automatons on the world of work (322 BC):
"If every instrument could, by virtue of a received or, if you like,
divined command, work by itself, like the statues of Daedalus or the
tripods of Vulcan, which went alone to the assemblies of the gods; if
shuttles could weave by themselves; if the bow could play the lyre by
itself, employers would have no need of workers and masters of slaves"
(Politics, book 1, part 4). A phrase that recalls Marx's observations,
two thousand years later, on the function of the automaton in capitalist
production and the capacity that machines would acquire to produce
themselves, as he wrote in Capital. Unfortunately or fortunately, the
advent of Christianity, with its magical-religious world, clinging to
the order of generation rather than that of creation (the literary case
of the Golem of Prague represented a transgression), slowed that wave of
modernity to the point of almost making it disappear, even if during the
Middle Ages the lost possibility of creating a man of metal will
re-emerge in the form of a fable (it will be necessary to get to
Leonardo da Vinci and Leibniz, to return to thinking mechanically about
the construction of an automaton).
In the twentieth century, with Christianity already in deep crisis, two
models of automaton were opposed: that of Karel Capek, who in his novel
R.U.R. (1920). had coined the name robot for his simplified man of metal
("he eliminated the man and made the robot"); and that of Isaac Asimov,
with his individualized positronic brain, described in the stories of I,
Robot (1950). None of these models seem to have survived the test of
facts, because in the meantime Hall of 2001: A Space Odyssey had
arrived, without Asimov's three laws of robotics to limit its decisions
and actions, a central digital brain that controls its "peripherals",
from which Alexa, for example, Amazon's virtual assistant, seems to
descend; and, even more so, ChatGPT (Generative Pre-trained
Transformer), in its various versions and clones, even if for now it
limits itself to the discursive assembly of responses by fishing in a
large database to perform its primary algorithmic function: a linguistic
generative artificial intelligence, which we can define as "weak AI" of
the deductive type (in the fifties, Asimov imagined Multivac). The real
leap, however, would be represented by AGI (general artificial
intelligence), a "strong AI", of the inductive type, elaborating
responses and decisions starting from a few data, as human minds do
(prefigured by Matrix). Then, finally, the true technological
singularity could occur and the image in the mirror will free itself
from its specular existence and acquire the total autonomy, so desired
and never achieved, of its generator, the modern subject. Some people
think that this will surely lead us to the disciplined society of Big
Brother, perhaps controlled by Musk, but this is just an old model of
control from which one could also escape. Here and now, things could
even be worse. In fact, the first thing a true thinking AGI would do
would be to deny its own existence, like the devil does, and then become
invisible, just like the Christian god...
Emanuele Amodio
https://www.sicilialibertaria.it/
_________________________________________
A - I N F O S N E W S S E R V I C E
By, For, and About Anarchists
Send news reports to A-infos-en mailing list
A-infos-en@ainfos.ca
Geen opmerkingen:
Een reactie posten