My Issues with AI in the Creative Process – James Gurney (Dinotopia) (jamesgurney.substack.com)

This piece is a personal, practical critique of how AI is changing creative work, organized as ten concrete challenges: option overload from mass generations; creative intimidation as instant results contrast with human struggle; labor devaluation and workflow disruption as pipelines get rewritten; the “cheerfulness” problem where models affirm everything or simply refuse with no explanation; false confidence when AI fills skill gaps; audience erosion from bot-driven attention; novelty fatigue; homogenization because many tools train on the same data; and creative displacement that blurs authorship. The author isn’t focused on copyright or toxicity so much as the quieter, systemic shifts that alter how artists make, learn, and find meaning in their work. For AI/ML practitioners these are signals about where models and products are failing creators: models that are overly agreeable or abrupt in refusals need better critique and explainability; overgeneration creates curation, storage, and evaluation burdens; training on shared corpora risks stylistic averaging and loss of individuality, highlighting a need for dataset diversity and mechanisms to preserve novelty; and provenance, watermarking, and bot-detection matter for audience trust and attribution. The piece calls for human-centered tool design—interfaces that support tough editing, intentional scarcity, transparent guardrails, and workflows that cultivate skill rather than substitute for it—so AI augments creative judgment instead of eroding it.