Generated Realities consists of two books constructed using AI-trained image models. Rather than producing contemporary "AI aesthetics," the models were directed to generate images that feel archival, degraded, low-resolution, plausibly recovered from early internet culture or forgotten digital cameras.

The first volume presents ambiguous domestic fragments: voyeuristic interiors, awkward intimacy, banal scenes rendered uneasy. The second volume shifts tonally. Pastels, glittering textures, and harsh flash photography evoke early Tumblr-era aesthetics. Fluids recur, hovering between decorative and bodily.

There are no margins or negative space; every spread is full-bleed. When two images share a spread, they repeat the same sentence. Language functions as echo rather than caption.