In recent years there has been growing interest in artificial neural networks (ANNs) which are quickly becoming the primary device for machine learning. Used for finding patterns in large data sets, ANNs were also recently employed in many artistic contexts: as tools for artists, semi-independent creators of content, and even as invisible “critics” which / who predict our aesthetic preferences. The aim of this paper is to speculate about the disruptive effect of these ‘alien agencies’ on the (modernist) aesthetic regime of art centred around the notion of autonomy. The author examines how neural networks and connectionist epistemologies may potentially affect the most common ways of producing, circulating, and valorising art. He claims that the possibility of automatizing creativity and art criticism may lead to the emergence of a new aesthetic regime based on forms of dynamic, distributed and probabilistic governance.