UT philosophy professor Y.J. Erden to give talk on AI "art" at Maastricht University on 1 April

From Generative AI we seem to have novel images from text prompts. Such work is sometimes technically good and can be visually appealing. We may even want to call these outputs ‘art’. They do however lack some rather fundamental as well as foundational features which are essential to artistic practice and experience. In this paper I outline what these purely digital outputs lack. I further argue that simply grouping these works with human derived art brings negative impacts to both art and artists, as well as to society more broadly. 

I group these concerns in three categories: (1) devaluation, of both the art itself and of the labour and skill of artists, (2) disruption, both to the processes of meaning-making and to essential experiences such as catharsis, and (3) dilution of what we understand by key aesthetic concepts like interpretation and authenticity. I end with a call to stop mis-categorising these outputs. Though not the primary aim, I hope this paper lends support to growing calls for greater regulation of the ways in which these models appropriate, and companies profit from, artistic material and data scraped from the internet.