RAFAEL MELENDEZ, MLII, Academy of Realness, 2025

Part of an ongoing sequence of artist publications made with generative AI, MLII is built from synthetic images in which art history is fed back through a machine and returned polished, recognisable and slightly wrong.

The source material is deliberately loaded. David Hammons reappears through Pissed Off — the 1981 action, photographed by Dawoud Bey, in which Hammons urinated on Richard Serra's T.W.U. — and through Bliz-aard Ball Sale (1983). Felix Gonzalez-Torres's strings of lights surface elsewhere. The front cover takes the sculptural language of Wade Guyton's U forms and returns it as an immaculate, almost commercial render.

The book belongs to Melendez's developing Psychofigurism, here operating as a kind of art-historical laundering. Existing images are not reproduced faithfully; they are passed through early generative AI, stripped of their original photographic conditions, and rebuilt from the machine's approximation of what it has learned.

The title performs the same operation. PLAGIARISM contains four letters that can also function as Roman numerals: M, L, I, I. Rearranged conventionally they become MLII — 1052. The word itself is processed through another system, leaving only what that system recognises.

There is an art-historical lineage through Sturtevant, whose repetitions of Warhol, Johns, Stella and others destabilised the distinction between original and copy decades before machine-generated imagery. AI introduces a further problem: it does not simply repeat the artwork. It reconstructs a statistical memory of it.

Sturtevant repeated the artwork. AI repeats the memory of the artwork. MLII works in the difference.

There is no conventional essay inside the book. The only text appears on the inside of the removable dust jacket, where AI is also credited as part of the publication's production. The text sits somewhere between manifesto, corrupted critical essay and machine confession: plagiarism becomes an "error state"; imitation becomes evidence of replication; archives generate "millions of digital ghosts"; and "the original disappears, replaced by circulation."

Its placement matters. The images occupy the book; the language explaining — and implicating — their method sits on its removable outer skin. Even the explanatory apparatus refuses a stable position.

MLII is not homage so much as reprocessing: art history treated as image matter, passed through AI and returned as synthetic residue.