← Back to A Note on AI & Submissions
When we ask of a submission, “What machinery produced this?”, artificial intelligence is only the newest machine in the room. Two older traditions already taught us how to read writing made by selection, arrangement, and rule rather than by invention alone. They are the inheritance a machine-assisted text steps into — and the reason we are slow to mistake a tool for a genre.
Found Language
The literary tradition contains many forms that blur the boundary between authorship and arrangement.
Found poetry reshapes existing language into new structures and meanings. The poet often works not by inventing words but by selecting, arranging, removing, reframing, and recombining them. The resulting work becomes a kind of linguistic collage.
This tradition predates artificial intelligence by many decades.
Writers have built poems from newspapers, legal transcripts, advertisements, speeches, instruction manuals, government reports, and fragments of existing texts. The act of authorship often occurs in the selection and arrangement itself.
From this perspective, prompting an AI can sometimes resemble working with found language.
The machine produces excess.
The writer selects.
The writer edits.
The writer rejects.
The writer reframes.
The writer discovers.
This does not make all AI-assisted work successful.
Neither does the existence of scissors make every collage interesting.
Oulipo and Constraint
The Means of Production is perhaps even more interested in another literary tradition: Oulipo.
Founded in 1960 by writers and mathematicians including Raymond Queneau and François Le Lionnais, Oulipo explored the use of constraints as engines of creativity. Rather than treating artistic freedom as the absence of rules, Oulipo treated rules themselves as generative devices. Writers intentionally imposed restrictions upon language in order to discover possibilities they would not otherwise have found.
A sonnet is a constraint.
A lipogram is a constraint.
A sestina is a constraint.
An erasure poem is a constraint.
A machine prompt can also be a constraint.
In this sense, AI may function less as an author than as a peculiar Oulipian device: a system that generates unexpected linguistic possibilities inside a set of formal conditions.
The question becomes:
What did the constraint reveal?
What became possible because of it?
What structures emerged?
What myths surfaced?
What hidden assumptions became visible?
These questions interest us far more than arguments about purity.
Both traditions point the same direction. The interesting question was never whether a machine was involved. It is what the selecting, arranging, and constraining made visible — and whether the result exhibits intelligence, surprise, pressure, or genuine strangeness.