31
Metafísica y Persona. Filosofía, conocimiento y vida
Año 15, Núm. 30, Julio-Diciembre, 2023, ISSN: 2007-9699
A natural articial intelligence? Some notes on the
computational biomimicry of human intelligence
¿Una inteligencia articial natural? Algunas notas sobre la
biomímesis computacional de la inteligencia humana
Centro de Bioética, Universidad del Desarrollo, Chile
hv_mx@yahoo.com.mx
This article introduces the case that it does not seem plausible that AI can come to
be presented as interchangeable with human intellect, as if its processes could pass as
natural, as much as our intellectual exercise of understanding reality is. The paper shows
that even though AI reproduces the structure of human knowledge yet misses subjec-
tivity. And in that sense, strong AI could not overcome human knowledge, because it is
not able to see itself as an active spectator of itself, nor protagonist or responsible for its
actions. Though some think that the human being’s own lies in a dynamic combination
interdependent sociability. It is of no interest to AI to imitate our biographical temporal
vulnerability, although it would be interested in imitating rational autonomy; and it does
not need interdependent sociability either.
Keywords
Este artículo introduce el caso de que no parece plausible que la IA pueda llegar a pre-
sentarse como intercambiable con el intelecto humano, como si sus procesos pudieran pa-
sar tan naturales como lo es nuestro ejercicio intelectual de comprensión de la realidad. El
documento muestra que, aunque la IA reproduce la estructura del conocimiento humano,
aún pierde la subjetividad. Y en ese sentido, la IA fuerte no podría superar al conocimien-
to humano, porque no es capaz de verse a sí misma como espectadora activa de sí misma,
ni protagonista ni responsable de sus actos. Aunque algunos piensan que la propia del ser
humano radica en una combinación dinámica de diferentes características como la corpo-
reidad vulnerable, la racionalidad autónoma y la sociabilidad interdependiente. A la IA
1
Metafísica y persona. Filosofía, conocimiento y vida
Año 15, Núm. 30, Julio-Diciembre, 2023, ISSN: 2007-9699
32
imitar la autonomía racional; y tampoco necesita la sociabilidad interdependiente.
Palabras clave
Introduction
-
sesses the principle of movement and rest in itself versus
tries to imitate the natural, sometimes it tries only to reproduce the natural,
but other times it tries to complete or perfect it.
refer to result.
achieve the synthesis of the active ingredient that triggers a natural reaction
-
with something natural, the limits of the natural and its potentialities estab-
they would not cut, but we can grow ears or aortic valves on the skin of a ro-
dent because for this it is only necessary to grow tissue within another tissue.
-
man cognitive abilities through its imitation, it focused its mimesis on the
operations of calculation and logical deduction, with the intention of making
encompasses all the advances that AI has introduced into our daily lives) and
33
A natural articial intelligence? Some notes on the
computational biomimicry of human intelligence
-
nition of what is meant by intelligence.
Articial intelligence strong and weak
AI refers, beyond an imitation of human calculation processes, the study
and imitation of the way in which human solve problems through processes
of symbolic information.
2
AI does not include non-symbolic numerical infor-
mation processing systems but implies the ability to address and solve prob-
AI awakened the expectation of reproducing intelligent processes according
were strategy games, and later resolution of logarithmic problems, logical rea-
soning, simultaneous translation of languages and automatic reading of texts,
3
great indisputable goals, especially in expert systems and its ability to make
-
ent criteria for information processing are used. Today expert systems have
multiplied their functionality: they are able to diagnose, monitor, plan and
interpret information. In other areas such as speech recognition, simultane-
sought. That is weak IA. But strong AI implies the development of intelligent
machines with scopes like or greater than those of human intelligence, but
without understanding the operations and processes it executes. AI could
pose as natural intelligence, but, to this day, hardware has only been able to
emulate cognitive structure, while software emulates the process.
4
From articial computing to natural AI
placed on the distance between AI and human intelligence rather than com-
2
(Eds.), Diccionario
Interdisciplinar Austral
3
A. (Ed.), Historia Universal del Pens-
amiento Filosóco
4
, G., Dizionario Interdisciplinare di Sci-
enza e Fede, Urbaniana University Press, 2002, vol. 1, pp. 771, 768.
Metafísica y persona. Filosofía, conocimiento y vida
Año 15, Núm. 30, Julio-Diciembre, 2023, ISSN: 2007-9699
34
us is the nature of the exercise of our intellect; that is, there is a qualitative
-
monstrative reasoning from universal mathematics and the assumption that
human rationality was but part of that mathematics, such that the automation
of calculus would only have to reproduce those rational processes to emulate
and eventually supplant natural intelligence. This suggested that any problem
or logical approach could be calculated or reduced to a computable algorithm.
With the miniaturization of electronics, logic machines emerged over
arithmetic calculators. With feedback and self-organized systems, computers
in their modern form made their way as logical-symbolic operators. Subse-
quently, expert systems surprised by their ability to emulate humans special-
ized in a certain subject and became able to solve all kinds of problems by jus-
tifying the resolution process: they managed to acquire information based on
theories and accumulated experiences. There were able not only to acquire,
process, and store information quickly, but also to select and discriminate the
the context of the problem and its elements and through the understanding
of the symbolic language. That is communication.
5
them, but from trial and error and the analysis of the particularity of each
case, the system was able to solve problems in a human way. That is, learn
and assume failures in the process of acquisition, management or increase of
knowledge, etc. Gödel’s theorem stated that there were truths that the ma-
chine could not reach while human intelligence yes could. But identifying
how IA imitate human does not mean having found a way to be intelligent
in a human way. The limitations can also occur in humans. It is therefore
is possible that it is imitated by AI so that it can seem natural. That is, if the
machine can really think and not just acquire, store and process information;
abilities in which they are indisputably superior to humans.
If we notice how the human neural network operates, we will notice that
it adjusts its semantics through learning: from children the human explores,
plays, rehearses, corrects by contrasting his body with the environment that
5
, T., Homo Cybersapiens. La inteligencia articial y lo humano, Pamplona: EUNSA,
2002, p. 39.
35
A natural articial intelligence? Some notes on the
computational biomimicry of human intelligence
surrounds him, facing another everything: it reveals a joyful, irritable, dissat-
-
pears, with its words and symbols, contact with reality is not carried out di-
rectly, through the external senses; neither only in a sentimental or emotional
way, but through formal signs and symbols, which only have the semantics
and feelings that we assign to them.
In that sense, both the human being and a computer manipulate symbols,
although the human knows that he only deals with signs and symbols. When
their action because if they tried they would put their survival at risk. In hu-
ideas we formulate on a particular issue are the right ones, and if not, what is
the way to adjust them.
-
ness of our ideas, we warn that information is only information. And when
that happens, the signs and symbols of language and communication are
shown as a reality that refers to something external (intentionality). When
the human being asks a question, he knows that the questions are only ques-
tions, that he will have to adjust, correct, or suspend them because he knows
that what has been achieved through the answers is not all the information
to be acquired, and this can only be achieved by an exercise of consciousness.
Machines and animals are intentional and require and acquire informa-
tion. While in the human intellect there is much more than that. In animal
intentionality the fundamental thing is to adapt to reality; he can only do
what comes to his mind, to ensure survival.
But human intelligence goes beyond that. It does not stop at conclusions.
If it only prompted answers and not new questions it would fail in an anthro-
pological sense because asking is a healthy sign of knowing how to put in pa-
ability by which we possess, master, or direct our information (and we know
that it is only information), is combined with curiosity that leads us to new
questions about each reality and the set of realities.
6
Human intelligence does not just solve problems: it creates them and gets
into more trouble; not because it wants to complicate existence, but as a way of
recognizing the complexity of the world around it in all its aspects. The human
6
, T., Homo Cybersapiens... pp. 133, 175.
Metafísica y persona. Filosofía, conocimiento y vida
Año 15, Núm. 30, Julio-Diciembre, 2023, ISSN: 2007-9699
36
carrying out projects, considering new possibilities, because when discovering
that ideas are only ideas, the human being warns that there is more reality than
thought. If AI did the same thing, it would have to be called natural.
and meaningful interactions with the intention of contributing, enriching, in-
creasing, or varying the understanding of real interactions.
In this sense, it does not seem that computer neural networks operate in
-
there is; they operate what there is and according to what there is, without
further questioning.
Rationality versus understanding
requiring thought: and AI seems to be primarily concerned with the imitation
The modes of reasoning, i.e., rational skills, are methods of knowledge,
but not of thinking. Asking questions about the known goes beyond ratio-
nal skill itself, which is what computer science and cybernetics all about are.
Asking myself questions about these methods of acquiring and processing
information, leads me to a meta-science, a meta-logic, a meta-mathematical,
which allows me to possess and direct reason (always instrumental) and its
results. Reason requires to be directed, while the intellect directs and allows
to act well or to run well. That is, to think wisely, sensibly, with the increase
in the capacity to understand what this entails.
-
and makes our intelligence ready to know and understand the world in front of it.
That turns our mind into a free intelligence that invents, dominates, har-
monizes, directs, rational skills, and turns them not only into skills for obtain-
ing information but into a source of understanding. Therein lies our pecu-
liarity, in our ability and interest to understand, not just to learn to conclude.
The social character intrinsic to our nature implies not only the mutual de-
pendence between humans to communicate but also for the mastery and en-
richment of our cognitive operations, which are not achieved in solitude and
37
A natural articial intelligence? Some notes on the
computational biomimicry of human intelligence
without dialogue, which is only achieved with the development of culture.
a man lies in our ability to cultivate dialogue as a basis for acquiring intellec-
tual habits that allow us to conquer the freedom of thought; because rational
machines do not actually possess the knowledge they imitate and reproduce,
nor do they have the freedom to know.
AI reproduces the structure of human knowledge but misses subjectivity.
And in that sense, strong AI could not overcome human knowledge, because
it is not able to see itself as an active spectator of itself, nor protagonist or
responsible for its actions.
Some think that the human being’s own lies in a dynamic combination of
-
ity, and interdependent sociability. It is of no interest to AI to imitate our bi-
ographical temporal vulnerability, although it would be interested in imitating
rational autonomy; and it does not need interdependent sociability either.
Conclusion. How natural can AI become?
It has been suggested throughout this text that it does not seem plausible
that AI can come to be presented as interchangeable with human intellect, as
if its processes could pass as natural, as much as our intellectual exercise of
understanding reality is.
-
ity of acquisition, processing, and increase of information, called rationality
(which would be imitable and widely surpass able by rational machines),
with respect to the intellectual skills that are exercised to understand reality,
not just go from one premise to another syllogistically, because AI does not
require understanding to compute, manage or optimize the information that
is generated in its processes.
7
Perhaps we have come to think about the possibility of strong AI by iden-
brain operability of the human being is carried out from the perspective of
the third person; while the most intimate of our mental operation, that which
is only self-transparent to the subject in which it occurs, belong to the per-
7
¿Qué tan natural es la inteligencia articial? Sobre los límites y alcances de la
biomímesis computacionalNaturaleza y libertad, núm. 12, 2019, pp. 245-256.
Metafísica y persona. Filosofía, conocimiento y vida
Año 15, Núm. 30, Julio-Diciembre, 2023, ISSN: 2007-9699
38
we notice that we know we exist, we know who we are and what we do. This
is not in the AI scenario because to work it does not require a biographical
singularity of the information processing it produces. The same can be said
others, we coexist and penetrate the intimacy of the other and manifest the
intimacy that constitutes us from within. This knowledge is reached from
the perspective of the second person.
When AI seeks to emulate, overcome, and supplant the unique character
of the human being, it reduces those two perspectives to one through the
naturalization of knowledge. It is a reductionism of human wealth from the
perspective of the third person.
And when we forget the individual, unique, unrepeatable character of
each of us, we are facing a reductionism of the second person, which is in-
tended to be possible in the latest developments of AI.
-
ously fruit of intellectual operability) is that we have blurred the scope, limits,
and realistic expectations of the variants of AI. Understanding this (not only
inferring it), calibrating it, weighing it, can help us avoid false reductionisms,
but also unwarranted fears. AI will continue to give us tools to not depend on
our limited discursive capacity, but it will still be at our disposal to be taken
in its true social, cultural, and technological dimension, and for this, there is
no biological mimesis that is enough.
References
(Eds.), Diccionario Interdisciplinar Austral
, T., Homo Cybersapiens. La inteligencia articial y lo humano, Pamplona:
EUNSA, 2002.
, G., Dizionario Interdisci-
plinare di Scienza e Fede, Urbaniana University Press, 2002, vol. 1, pp. 767-781.
A. (Ed.), Historia Univer-
sal del Pensamiento Filosóco
Naturaleza y libertad, núm. 12,
2019, pp. 245-256.
