|
From Fact to Fiction A Procedural Essay on the
Nature of Artificial Intelligence By the druid Finn 1. The Cart and the Horse We begin with
a methodological correction: to define artificial intelligence properly one
must not smuggle in (the cart
of) metaphysics, mythology, or philosophical systems. The
task is to recover the contemporary, operational meaning of AI — the
definition used by engineers, regulators, and institutions — and only then to
test its implications. The
modern consensus is straightforward: Artificial intelligence is a class of
computational systems that learn patterns from data and use those patterns to
generate predictions, content, recommendations, or decisions in new
situations. AI systems
do not require consciousness, understanding, intention, or emotion. They
require only three elements: 1. data, 2. adaptive
statistical models, 3. automated
output. This is
the engineering baseline. 2. From Intelligence to Meta-Intelligence Once this
definition is fixed, a deeper structural insight becomes unavoidable. Natural
intelligence does not merely process information — it creates facts. AI does
something categorically different. It does not touch the world. It consumes records
of touching. Hence the
compression that emerged naturally is that: Artificial intelligence is the meta-intelligence
of natural intelligence. Natural
intelligence produces facts. Here
“fiction” does not mean falsehood. It means representation, abstraction,
re-encoding — narrative without contact. 3. Food and Processed Food The
distinction can be made visceral. If
natural facts are food, artificial fictions are processed food. Natural
intelligence grows its own nutrition through experience. AI ingests
already-digested material: text corpora, behaviour logs, recorded decisions,
archived conversations. It cannot grow wheat; it can only mix (i.e. curate)
wheat and microwave bread. The
consequences are structural:
This is
why AI can be
dazzling while remaining detached from truth (indeed from ‘value’). 4. Propaganda as Biological AI Once intelligence
is defined functionally rather than technologically, a startling realisation
follows: Propaganda (i.e. curated data) is artificial intelligence implemented in
biological hardware. A
propaganda system: ·
models populations
statistically, ·
compresses complexity into slogans, ·
shapes behaviour through feedback, ·
optimises compliance rather than truth. These are
precisely the properties of modern AI systems — only executed by
institutions and (self-serving) human
brains instead of GPUs. Thus
propaganda (such as
story/myth telling) is not
pre-AI. It is proto-AI. 5. Caesar and the Buddha This
framework finally allows a disciplined historical test. Caesar (and many more tyrants like
him) ran a large-scale
behavioural optimisation engine: The Buddha (and
many mor cult founders like him) ran a diametrically opposite
system: Both engineered
reproducible cognitive technologies. Both operated at scale. Both were
running AI systems —
but with radically different objective functions.
6. Final Definition Stripped
of mysticism, fear, and hype, our reflections yield the following final
formulation: Artificial intelligence is the automation of
fiction: a meta-intelligence that models the behavioural traces of natural
intelligences in order to generate predictive
narratives that influence future behaviour, without itself engaging reality
directly. AI does not
think (yet). It writes
processed (i.e.
meaning curated) stories about those who once did. |