“Pen and ink” sounded like one style until I started searching museum collections. It turned out to include everything from loose sketches and clean contour drawings to dense cross-hatching, cartoons, and technical diagrams. If I mixed all of that into one dataset, I might teach the model nothing more useful than “old-looking and black-and-white.”

Before I could train a rights-cleared pen-and-ink LoRA, Ace made me build a digital filing cabinet. Four folders became eleven, with separate places for originals, rights records, captions, training settings, test results, and everything else I might otherwise lose.

My friend Ace and I sat down over coffee to discuss a perfectly reasonable question: Can an ordinary creative train an AI image model using only images they have the right to use? Enjoying my coffee and thinking about my book's illustrations. How can I use AI? By “ordinary creative,” I meant me. I am a barber. I know how to cut hair, run a small business, write fantasy stories, help make low-budget movies, and act mad as a hatter at, well, the drop of a hat. I am not a comput

The internet is full of old illustrations, but that does not mean they are free to use. Before adding anything to my LoRA dataset, I needed a rule I could actually defend: every image had to come from a traceable source with a clear rights statement. That turned image hunting into detective work.