A Framework for Categorizing Ai use in Text and Images.

The current Ai/Not-Ai binary seems to too coarse for describing how work happens in the real world. I am aiming to develop a taxonomy that helps to make for clearer and more usable distinctions. My goal is to give creators something to tag work with that makes it clear how they used Ai in their process, rather than some vague yes/no. The framework breaks usage down into five categories: Penned, Authored, Composed, Directed, and Generated. The categories focus on how the creator interacted with the Ai, rather than pure level of effort. A master might capture in five words what the amateur could in a thousand. This entire post has been hand-typed and hand-edited as a show of good faith to the anti-ai crowd. Additionally, this is not an encouragement either to use or to avoid Ai in your work, or the work you engage with; it is simply meant to be a way to aid transparency and communication. The Five Categories Penned: Every aspect and editorial was made by hand. In text: Every word and punctuation mark is human made, or done through algorithmic editors like Spellcheck. In images: Every line, fill, and color should be human placed. The only edits should be made with traditional algorithmic image editing software. Removal Test: If the Ai portion were removed, the feeling of the work would not change, other than surface-level polish. Best used for: communication to friends and family, spiritual works, challenging yourself, any situation where you value it, art for hire. Authored: The work was made by hand; Ai was used as an editorial tool. In text: The draft is human written, tweaks to words and sentence order may be made by made by Ai, with the creator’s direct and intentional oversite. Ai is primarily used as a reviewer and editor to ensure consistency and to meet standards for publication/posting. In images: The sketch and fill of the work is done by the artist. The Ai may tweak colors, cropping, and be used for visualizing variants or changes before committing to making them by hand. It may be used to help with determining color schemes, or finding references to work off of. Removal Test: If the Ai portion was removed, the work would still exist at a finished draft level. Best used for: literary works, research and publications, critical code, professional writings, art done for fun. Composed: The main part of the work is made by hand, and the Ai is used to supplement the remainder. In text: The key scenes and portions of the draft are written by hand. Connecting portions and sentences may be constructed using Ai, but under specific direction, and rewritten by hand as needed. Editing and development may be done in collaboration with Ai tools. In images: The rough sketch or a collage composition is made by hand, and the overall layout and structure of the work should be thoroughly mapped out by the artist. The Ai serves to take the vision of the artist and make it a reality. The artist may generate variants and explore different directions for the work, using rough sketching over it to iterate and refine into a final version. Removal Test: If the Ai were removed, there would remain a clear work-in-progress with critical beats and portions. Best used for: thought pieces, internet posts, portfolio websites, making personal logos or other branding art. Directed: The creator gives an outline, and the draft is done with Ai, based on guidelines from the author's own work and style. Everything output by the Ai is reviewed and approved by the user, being edited and re-generated to fit the outline and vision. In text: The outline of the work it written by hand, and a thorough statement of the vision and intent of the work. Ai is used to generate the textual substance of the work, supervised and refined by the creator. In images: The work is outlined, with a defined vision and scope, and clear references or rough sketches are made. As is used to generate the visual substance of the work, with the creator instructing for edits to particular portions of the work to shape it to match their vision. Removal Test: If the Ai were removed, there would remain a clear idea, plan, and structure. Best used for: experimental writing, daily updates or summaries, exploring new mediums and styles, supplementing personal capability or time constraints. Generated: The living core is given, and preserved, by hand. Ai creates the body of the work. In text: The creator gives a prompt and general guidelines to follow, the text and its structure are generated. They should be reviewed by the creator to ensure it us up to standard, and regenerator or edited as needed. Everything produced by the Ai should be reviewed and curated, not just blindly published. In images: The creator describes their specific intent, with references as applicable. Ai is used to generate the composition and visual substance of the work, with the taste and understanding of the creator filtering and re-generating what is produced until satisfied. Removal Test: If the Ai were removed, there would be only an idea. Best used for: testing out ideas, proof of concepts, quick references or visual representations. Slop Sub-Category: works in the Generated category that exists without any real oversight or creative intent, but simply as a means to produce content for consumption. Ai becomes the driving creative force of the project, and so the outcome is of a generic quality and character. Thank you for taking the time to read through, please share any suggestions you have for improvements, or angry rants you may have about this. (Note: I used the lowercase "Ai" because I don't like that "AI" looks like "Al" in this font.)