Randomness and recommendations – exploring the web platform of the artist Ivar Arosenius and other digital collections of art

Abstract

This article concerns efforts to increase the accessibility of art through digital collections on the web. Features for display and comparison used by the platform of the Swedish artist Ivar Arosenius and by web projects such as Europeana and Google Art & Culture will be discussed. These features include random display, recommendations and machine learning of visual similarities. Key questions are: What are the general implications of these features for the art history canon? What are the relations between the digital collections on the web and the physical museum? There is a risk that recommendations based on machine learning will come to influence the way we perceive similarity. If the canon is to be challenged, a closer cooperation between the digital collections and the physical exhibitions will be needed.