/imagine represents a series of virtual images
generated with the help of Artificial Intelligence. It focuses on an issue that
most AI systems have nowadays: creating human hands. The title of the artwork references
a typical text prompt that AI platforms use to activate the text field for
human to describe what the system should generate next.
With the recent leaps made in the visual domain
by AI platforms, the project also wishes to raise awareness of using AI as a
tool for artists in general. Is AI just another tool in the “hand” of artist?
Or can it ultimately accomplish new visions by itself?
Since AI has not yet a real understanding of
the meaning and cultural concepts behind the images, what is the role of the
artist using AI as a tool to guide the creation process? How much does the
human creator needs to steer the AI, by generating variations until a desired
result is achieved?
Other topics that need to be answered are related
to the copyright of the artworks. Since large data sets of existing images or
artworks were employed to train the AI models, is the result a truly original
vision or is it just a synthesis of many existing pictures, a glorified
statistics?
The current project aimed at discovering and unraveling the topic, by hands-on experimenting with the AI tools. The general
theme and vision for the artwork was identified by the artist, selected from current
issues of the technology. The actual process of working with AI platforms to
generate the hands still involved a large amount of time investment. Many
trials and experiments were performed in order to be able to select most
expressive hands. Not often the vision that an artist has in mind is
immediately reproduced by imputing text in the prompt. Since AI still has
limitations on the symbolic behind hand gestures, most of the time it was
rather challenging to achieve even basic hand gesture using the text input.
On the other hand, AI proved to be rather surprising
at times, opening new directions for the artist that were not always very clear
from the beginning. In the end, as creators, we found that the results exceeded
our expectations, the selected images have a certain artistic appeal, both on a
visual language level, as well as of the expressiveness of the gestures.
Despite the current limitations of AI, we are awaiting the (r)evolution in the
field and see how to integrate new tools in our creative process.