How Do You Keep AI-Assisted Art From Losing Its Original Identity?
retained text
I've been working with ChatGPT on a personal creative project: an original post-apocalyptic story called *The Stranger's Callsign*. Recently, I experimented with turning one of its visual concepts into a short cinematic video. The process was interesting: starting with an image, then adding camera movement, lighting effects, transitions and atmospheric radio sounds. But I ran into an interesting creative challenge. Post-apocalyptic settings have a lot of familiar visual elements: abandoned buildings, fog, radio towers, dark industrial environments. AI can reproduce these elements quite easily, but that doesn't automatically make the result feel original. I don't want my fictional world to look like a copy of an existing game or movie. It needs its own visual identity, history, locations, characters and atmosphere. The story and its characters are mine. I use AI to help explore and visualise the world, but the creative decisions remain mine. I'm curious about other people's experience: **How do you guide ChatGPT or other AI tools toward something original instead of getting another familiar-looking genre image?** Do you work with visual reference guides, detailed world descriptions, specific visual restrictions, or something else?
document fetched 2026-11-15 00:19 · canonical https://reddit.com/r/ChatGPT/comments/1xw25gx/how_do_you_keep_aiassisted_art_from_losing_its