/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 unravelling 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)evolutions in the field and see how to integrate new
tools in our creative process.
