Page 350 - The Black Box Book

Intelligence does not have to be human in order to be intelligent. However, A SERIES OF IDIOTIC DECISIONS
digital technology, including AI, might be viewed as a prosthetic limb that we
never had. (Fabuš, P., 2019, pp. 75–91). The capabilities of AI can then be viewed
as those of a human who is highly superior only in a relatively narrow domain.
“In our project, the provocative word Idiot plays a role in the basic vulgar
diagnosis of a possible relationship between the aforementioned agents.
Since the clinically outdated diagnosis Idiot Savant captures very well
the capabilities of current learning algorithms, the vulgar and offensive
connotations are taken from spoken language to strongly emphasize
the hidden extraction mechanisms. As described by McKenzie Wark in her
book Capital Is Dead: Is This Something Worst? the vulgarity of spoken
language is probably one of the few artistic strategies that can clearly
highlight the written – machine-executed – language of learning algorithms
(i.e., vectors)” (Metazoa-Org, & Javůrek, T., 2021).
Importantly, aesthetic objects in the context of learning algorithms can
be viewed not just as their outputs, but also as the forms required for learning
them (databases, interpretive frameworks, parsing, tagging, categories, etc.).
Seeing these objects in the context of artistic production is not automatic.
Their direct representation is practically non-existent, and we must resort to
a not so appealing visualization, which is frequently furnished by graphs and
explanatory words. A further obstacle is that creating and reading such graphs
and texts requires interdisciplinary knowledge and experience. Nonetheless,
there are a growing number of perceptive authors and teams who can effectively
unpack these objects, Including, but not limited to, Vladan Joler, Trevor Paglen,
Lev Manovich, Joanna Zylinska, and numerous others. However, the difficulty
of working with these objects persists because we have a hard time identifying
artistic expression, artistic strategy, shorthand, gesture, or style in them, which
are attributes we consider to be fundamental to art.
In his text Objects in Mirror are Closer than They Appear, Timothy Morton
mentions Socrates’ conception of art as something that is not a representation
but a display of inner demonic forces to illustrate that art is not about inter-
pretation or capture but rather attunement to the inhuman. (Morton, T., 2013,
pp. 15–39). This view is very similar to what we propose as the enabling frame-
work for an artistic perspective that reflects learning networks and captures our
working process in the exhibition AI: All Idiots. Artificial intelligence is a complex
tool that we as artists learn to play, that is, to create art, by tuning in. The process
of tuning, according to Morton, involves improvisation, which is a juxtaposition
of reading and writing.
350
To demonstrate the relationship between art and AI as closely as possible in
accordance with artistic practice, we chose to adopt a strategy similar to that
advocated by Morton, namely “tuning in” to art and artificial intelligence. In anal-
ogy to the nature of the decisions made automatically by algorithms and their
programmers, we have focused on making the individual actions and decisions
we make within the project almost slavishly and ‘idiotically’ straightforward,
purposeful, seemingly logical, or determined by the capabilities of the available
technology, i.e. those that appear neutral, invisible. In line with the attempt
to approach the character of artificial intelligence, we have decided to invert
the role of man and machine. Artists have “tuned in” to the logic of the dataset
and produced their own outputs, which can be seen as the work of production
that we usually attribute to tools or machines. As unambiguous as the connection
between the artistic and production roles may seem, the process of making art
is in many ways similar to the process of algorithm performance.
By assigning a production role to the artists, we also wanted to capture and
identify the automated, and therefore seemingly neutral, steps in the process
of material collection and network learning as significantly creative. In this way,
we were able to show the artistic practices that underlie learning algorithms
and offer a vision of what “other” knowledge artificial intelligence provides.
To live up to this, we have tried, at least to some extent, to let go of the artistic
and human expectations associated with the results of our work. The individual
results on display in the exhibition then show the limits of our ability to avoid
these expectations.
You could say that the first idiotic decision in the project was that
the artificial intelligence related to art would be trained on a dataset of artworks.
And since the exhibition is intended for a domestic audience, we targeted
the field of contemporary Czech art directly, and therefore Czech artists. But
the very next and key question is, who is the contemporary Czech artist? In
many cases, the status of the artist is obvious. Many other people have studied
art, have their own web presence, but do not pursue it professionally. Some
consider themselves artists, but the art world does not reflect that. Others
believe that what resonates in the media today will be forgotten tomorrow and
artworks that are not currently seen in major galleries will be written into art
history. There are many folk artists who are popular, but galleries do not exhibit
them. Artists working with traditional art techniques, on the other hand, tend to
look down on conceptual artists and digital art. And who is a Czech artist? Is it
one who works in the Czech Republic or one who was born here? Who is and is
not a contemporary artist? There are a number of problems. And it goes without
saying that the definition of an artist is very blurred. Moreover, the question of
who is and is not an artist is linked to an even more complex question: what is
and is not art? This question is answered not only by each artist, curator, gallery
critic and viewer, but also by each work of art on its own. What is and is not art
is, moreover, one of the questions that should remain open.
We found the answer in line with engineering practices applied to complex
problems, which we can also encounter in the inner workings of learning algo-
351