AI Haiku 9
Video, 2026 · Unique
Reserve Ξ0.03Buy Now Ξ0.10
Loading auction…
About the artwork
Ai Haiku are video poems built on haiku's principle of compression, where a single image or gesture occupies the screen just long enough to register before the piece lets it go. Each one isolates a private experience, something usually too quiet to count as subject matter on its own, and holds it under brief, exact attention.
A face that begins as a single, recognizable presence gradually loses its coherence, frame by frame, until what remains on screen no longer holds together as one person. The unraveling proceeds quietly, without spectacle, the way attention itself sometimes drifts away from a stable sense of who is doing the looking. The piece treats this dissolution as an ordinary event, something that simply occurs and asks nothing of the viewer beyond watching it happen.
Myself
Audio explanation
Listen to Francesco D'Isa discuss this work.
Francesco D’Isa (b. 1980, Florence) is a philosopher and digital artist whose work explores the aesthetic and socio-technical implications of artificial intelligence in contemporary visual culture. With a background in philosophy and a long-standing engagement with digital media, D’Isa approaches AI as a semiotic and conceptual partner within a distributed creative process.
His practice draws on a sustained investigation of the interplay between language, image, and meaning, using text-to-image models as a space for visual experimentation. Through refined prompt engineering and a critical engagement with the machine’s latent structures, D’Isa constructs images that reflect personal poetics while activating the shared ideograms encoded in our collective visual memory. These works traverse the latent space of generative models like a symbolic landscape, negotiating the boundary between authorial intention and algorithmic emergence.
As a theorist, D’Isa has contributed significantly to the discourse on AI and art, most notably in The Algorithmic Revolution of Images (Sossella Editore, 2024), where he examines the philosophical stakes of image-making in the age of generative systems. Rejecting the romantic ideal of the solitary genius, he advocates for a model of creativity that is relational, historical, and technologically mediated. His view of AI as a co-author—not anthropomorphized, but understood as an apparatus shaped by cultural and statistical memory—echoes the perspectives of thinkers like Flusser, Latour, and Glăveanu.
His recent works, including Sunyata (Eris Edizioni, 2023), the first AI-assisted graphic novel published in Italy, explore themes of impermanence, paradox, and the instability of representation. Often embracing the “errors” of generative systems, his images find poetic value in algorithmic drift and failure.
Central to D’Isa’s approach is the concept of semantic attractors: words or phrases that act as gravitational centers in the latent space of AI models, pulling visual output toward dense clusters of cultural meaning. By working with and against these attractors, D’Isa reveals how images are shaped not only by prompts but by deep, encoded patterns of interpretation, offering a vision of AI art as both a conceptual instrument and a mirror of our symbolic systems.