Could the interactive installation DEEP ALICE be a dreaming machine? Could it actually dream? After all, “who do you think dreamed it?” Alice asks at the end of Through the Looking-Glass. Drawing from fragments of illustrations from this book and from Alice’s Adventures in Wonderland by Lewis Carroll, visitors are invited to construct scenes that feed a machinic “imaginary”: a machine capable of capturing the produced scene and recreating it through an instant image-generation system using Artificial Intelligence.
To this end, the machine automatically generates text fragments from the two books and uses them as *prompts* (phrases that influence, or even determine, the content of the generated image). These prompts grow in complexity as the visitor interacts with the installation. In association with the scenes created by visitors, these prompts produce images that become increasingly nonsensical and incoherent from a human perspective. Yet from a machinic point of view, however random the results may seem, they remain consistent with the system’s internal combinatorial and computational logic.
In DEEP ALICE, the depths of computational processes are made tangible through improbable images, though entirely logical in their essence.
