Page 270 - The Black Box Book

In any case, the Parisian artist collective asked Barrat if they could
use ‘his’ algorithm for their ambition of democratizing AI art. He agreed
and even helped the group with the technical implementation of their
project. However, when the portrait produced with the GAN technology
was finally auctioned off at Christie’s, the 19-year-old was very upset. “Am
I crazy for thinking that they really just used my network and are selling
the results?”, Barrat publicly vented his annoyance via Twitter. In response
to the massive criticism, one member of the group replied, “We are
the people who decided to do this,” … “who decided to print it on canvas,
sign it as a mathematical formula, put it in a gold frame” (quoted in Bogost,
2019, online). Media scholar and game researcher Ian Bogost commented
on this statement in an article for The Atlantic magazine with a quote by
Andy Warhol: “Art is what you can get away with” (Bogost, 2019). But it
is not that simple after all. Even if the big scandal around the portrait
ultimately remained absent, there was still an intense debate about
who is the creator and author of the image. For example, in an article of
the German newspaper Süddeutsche Zeitung, Bernd Graff put forward
the thesis that it was not the portrait but the algorithm itself that should
be considered as art (Graff, 2019, online). The sarcastic comment of
a humanist could be: At least these algorithms are still being developed
by humans and not yet by other machines.
Indeed, it seems more important than ever to emphasize that AI sys-
tems operate not fully autonomously or automatically when they produce
artifacts that pass for art or when they perform activities that are consid-
ered creative. Humans are always significantly involved in these AI-based
processes in all phases of development: from the creation or compilation
of learning data to the design and application of an algorithm suitable for
the problem to the ongoing training of the AI model (Ernst et al., 2019).
At least in the near future this is unlikely to change, even though it
has long been a reality that machines interact with machines in order to
be able to optimize themselves. However, it is not only for this reason why
in the debate about creative AI it is simply insufficient to again emphasize
the contribution of human action in comparison with the supposedly auton-
omously performing machine. Instead, it would make more sense to aban-
don the schematic contrast and isolated view of the relationship between
humans and machines and to focus on the socio-technical prerequisites
for the production and formation of AI art, i.e. the specific contexts of
the infrastructuring and networking of practices, knowledge and media.
Such a decentralization of the view on human-computer interaction is
obviously very difficult for us, because concepts such as art and creativity
are deeply anthropocentric or anthropological. We are so used to think
of creative acts as results of a certain will, or to see art deeply entangled
with human experiences such as joy, self-doubt, astonishment, and de-
pending on social dynamics between artist and audience (Peters, 2021,
p. 11). The list of further distinguishing features could easily be continued

and differentiated, which is why we cannot simply get rid of them. Doing
so, the reference to the emotional-affective aspects of human creativity
270 271