By Katrina Geddes.
Today, almost anyone can create a “new” work in the style of a popular artist by prompting a generative AI model. The mimetic capabilities of generative AI are making it harder for artists to control their personal brands and for consumers to locate authentic works. This Article provides a novel taxonomy of the attribution interests implicated by AI’s mimetic capabilities. It distinguishes between “misattribution” (when AI-generated art is falsely ascribed to the authors of copyrighted training data) and “non-attribution” (when AI-generated art fails to acknowledge its causal origins and thereby obscures the labor of human artists). Proper attribution can mitigate the substitution effects of generative AI by protecting authors’ reputational capital and preserving markets for authentic goods. This Article offers a roadmap for strengthening the limited protection provided to attribution interests under existing law.