Operative ekphrasis: the collapse of the text/image distinction in multimodal AI

Abstract

This article discusses the implications of multimodal artificial intelligence (AI), including image generators such as DALL·E, for the traditional concept of ekphrasis. Using ekphrasis as an example of ‘thinking with AI’, it takes up the suggestion that in the digital realm ekphrastic relationships should be understood as performative rather than representational. Since with the introduction of modern AI the digital realm needs to be divided into a sequential part (classic algorithms) and a connectionist part (artificial neural networks), the article shows how the latter part ultimately tends toward a collapse of the text/image distinction in the technical system. Artificial neural networks both encode images and text as the same type of information, and they do so differently from the sequential model. Only in the context of multimodal AI, unlike in analogue or sequential paradigms, ekphrasis goes beyond the separation of or transition between text and image, but rather transcends this difference.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 23 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.