ABSTRACT
This chapter examines the role of prompting and takes a wide view of the technological, cultural, and political forces shaping cultural practices surrounding visual GenAI. The use of the verb ‘to prompt’ has taken on a new layer of meaning reflecting a fundamental shift in user interactivity. Text-to-image prompting performs a near-instant structuring of data with language inputs to generate intelligible images. We argue that prompting is not simply a new means of interaction with visual GenAI but a disruptive new cultural practice, part human intention and part machine automation, that lies at the centre of a third shift in cultural production within modernity. This occurs within a historical context after the previous major shifts of ‘mechanical reproducibility’ and then ‘digital remix.’ In the context of these shifts in cultural production, we consider key conditions characteristic of prompting. Our analysis explores the black box and the problem of bias in training data and in the prompter, and, within a bigger picture, critical prompting as a negotiation with the economic and political framework, an accelerated and algorithmically data-driven outsourcing of human choice, that structures and sustains both visual GenAI systems and the cultural practice of prompting.
