Page 24 - The Black Box Book

works cannot be limited to the preservation of the work’s material as-
pects in the sense of obsolete technology. The author also emphasizes
the issue of the isolation of online archives and museum collection por-
tals, both in connection to the general public and between themselves. In
this regard, he is convinced that the ideas of social networking must be
integrated into the archival practice of digital art. Accordingly, digital cu-
ration and preservation tools should encourage critical reading, qualitative
approaches, paratextual apparatus, and broad performative interaction
and conversation as guiding principles for cultural platforms. Moreover,
the author does not address the question of an appropriate and optimally
functioning technical infrastructure of online archives in isolation but rather
in the context of an ecological discussion, understanding media ecology
as a condition and accompanying phenomenon of natural ecology.
Section AI Curating is dedicated to artificial intelligence (machine
learning) as a new phenomenon in the field of digital curating. AI is pre-
sented through its applications in the field of art, ranging from the gen-
eration of images based on pattern recognition in databases of human
creations to AI as a curator of extensive digital collections of antiquated
art. Furnished with the capabilities of accelerating and extending hu-
man abilities to work with big data in art history, we can understand AI
as a new player in contemporary digital curating. Also, AI is discussed as
an object of artistic research, critically reflecting on AI’s social and episte-
mological function. The novelty of this issue has influenced the character of
the majority of contributions in this section, which tend to focus on detail-
ing specific projects rather than offering cursory observations on the topic.
The author of the chapter On Computer Creativity. Machine Learning
and the Arts of Artificial Intelligences describes how artificial intelligence
works in generative arts. Andreas Sudmann contends that AI systems
serve as assistants for people rather than taking the lead in the creative
process. He argues that human involvement in these AI processes is
always significant in all phases of development, from the collection of
learning data, construction of models, and the fine-tuning of algorithms
that are the basis for training the models. It also critically examines
the criteria used to evaluate AI creativity and challenges why we typical-
ly settle for an AI model’s capacity to imitate human creativity. Central
to his argument is defining and redefining creativity and art in relation
to AI-generated works. Here, a space is opened for discussion about com-
puter creativity that should also challenge us to take the contradictions of
human creativity and art production seriously. In light of the always-chang-
ing dynamics of artistic production and creativity within the contexts of
industrial and post-industrial societies, he asks how machine-like or how
dissimilar human art and creativity are from AI.
Lukáš Pilka recounts his personal initiative, in which he deployed
AI (machine learning) as a curator of digitized collections of antiquated
art in a network of Central European memory institutions, in the chapter
Digital Curator in the Museum of Fine Arts. The AI tools he created for
this aim were used to look for works based on artistic techniques, sym-
24
bols and taste. Users can curate their selections from the database using
the Digital Curator, a customized website he developed based on recur-
rent motifs characteristic of a particular period in art. The intention was
to respond to concerns raised by AI research on knowledge extraction
from art collection databases, where software developed for the visual-
ity of contemporary popular culture is frequently used to analyze works
from earlier historical periods. To verify that the AI does not, for instance,
confuse a halo with a hat or identify an angel in a painting, Pilka’s team
manually edited the dataset that has been given the proper metadata.
Similarly, the chapter The Next Biennial Should Be Curated by a Ma-
chine details a specific curatorial effort that infuses AI into the creative
process in the capacity of an exhibition curator. The creators of this AI
software, UBERMORGEN, explain how this tool was utilized to create
the exhibition concept after the AI was taught the know-how of
the curators of the Liverpool Biennial shows. In addition to placing AI in
the place of the object described in this chapter, the author also incorpo-
rates AI into the translation and editing of text by using AI tools. As a result,
the chapter’s reader gains a quick understanding of the kinds of effects
that the collaboration between humans and AI can produce.
AI: All Idiots is the title of the final chapter in the section dedicated
to AI curating. Here, Barbora Trnková describes the collective exhibition of
the same name, whose creators critically tested various facets and phases
of the machine learning process: from manipulation of input data during
dataset processing to the biases introduced in the process of automat-
ed cognition and the technical limitations of the tool itself. It challenges
the premise that AI is incapable of providing novel creative solutions
and draws out the tension between creativity and the perpetuation of
the status quo. The artworks on display serve as both testament to
the conflict between human artists and the AI tools’ machine logic and
further as evidence that creative AI research has the potential to be
a distinctive type of critical reflection addressing this unravelling medium.
25