Machine Learning in Artistic Practice#
Developments in Artificial Intelligence are changing many areas of society and life. Accordingly, it is important to reflect and shape them from different perspectives. Art can enable an independent access to this abstract technology and thus make a relevant contribution to the coexistence of humans and machines.
In everyday use, algorithmic processes are hardly perceived. In art, they appear in sensual forms and outside their original functional contexts. This enables us to sharpen our perception of this technology and the artefacts it generates.
Nevertheless, the examination of AI through art usually requires rational knowledge. Often, the process of creation, i.e. the technology, is an elementary part of the work. This must therefore be understood in order to be reflected on the basis of the artwork.
Typical forms of the artistic use of AI are the generation of new material or the classification of existing material; the results of the calculations are transferred into fixed objects or are constantly changing; the works thematize the abstract internal structure and/or the concrete objects outside the neural networks. The works are often created in an iterative process in which code and concrete object are adapted to each other.
Exemplary approaches: Much of the current success in machine learning is based on large data sets. The conditions under which they were created, the assumptions depicted therein, and the relationship between a concrete sensual object and its representation in abstract symbols are thematized artistically. The cultural technique of classification is criticized. Algorithmic processes, especially neural networks, create a space of fiction. Excerpts of it become reality, sometimes with a considerable amount of speculation. How do we deal with these artefacts?
The talk was held in slight variations on the following occasions: