GAN / Michael Friesen
This Artwork Does Not Exist

An image produced by a form of Artificial Intelligence, GAN (Generative Adversarial Network), trained by Michael Friesen on images of modern art.

The relevance is quite conceptual here – partially, it may be related to my Artificial Ignorance Choir, as a means to create tension between intelligence and ignorance when it comes to machine learning. On what grounds can the one or the other be ascribed?

Another aspect is intentionality. One of the central qualities of – or prerequisite for – an artwork in many theoretical discussions is intentionality of its creation and/or designation as art by its maker. That would be typically the grounds for exclusion from the category of art of any natural phenomena regardless of their aesthetic qualities. To what extent can a machine intend then? Can goal-orientation (even if pre-programmed by somebody else) be equated to intention? And whose intention is it then – machine’s or it’s programmer’s? And if it is programmer’s intent, but the programmer is not considering themselves an artist – is an image that resembles an artwork to be considered an artwork? Would that assessment change if the image is generated on the basis of other images that are canonised as artworks? Introduction of the machine in this circuit enables destabilisation of many concepts taken for granted. That can be useful as a reference in some of my projects where I wish to posit as my artwork something that was not generated by myself directly, but by my audience in response to something I have presented for them. Not that it is very important for me to claim that authorship, but rather as the means to redirect attention to that derivative product as the more important output from my process than any particular thing I may have assembled by my own hands.

Another interesting aspect is the work’s existence solely as a digital image – something that I have explored through my pdf-exhibitions.