The Inevitable Solution to the AI Art Dilemma (medium.com)

This essay reframes the AI art debate away from “Art as Process” vs “Art as Product” and toward “Art as Language” — a mode of communication where meaning is conveyed indirectly and requires both technical skill and fine-grained control to evoke subjective nuance. Drawing on deconstruction, the author argues that complaints about AI art’s missing “soul” are primarily complaints about lost control: large generative models synthesize millions of existing works, and no amount of prompting can guarantee the artist-level decisions that make a piece personally expressive. The result is a tension between AI’s promise to lower production effort and the artist’s need for precise control over what gets communicated. Technically, the essay urges the AI/ML community to move beyond “consume-all” training regimes toward modular, artist-centered approaches: small curated corpora, fine-tuning or plugin-style modules that encode an artist’s specific methods, composable algorithms, and toolchains that let creators control every stage of generation. It acknowledges prompt engineering and emerging tooling as partial fixes but calls for research into interpretable, controllable generative architectures, provenance and consent-aware datasets, and pipelines that let artists programmatically repurpose personal materials. The practical implication is clear: meaningful AI art will come not from bigger indiscriminate models, but from systems that restore artist agency and enable precise, communicative control.