What's in a Name? The Impact of Artist Names in Image Generative AI Prompts

Generative AI models have put image generation in the hands of the masses. While some contend this as a boon for creativity, others have argued about the legality of using such tools for commercial purposes. Such models rely on large training datasets that consist of images scraped from the Internet, many of which are copyrighted. These models enable the replication of visual styles of individual artists or brands who have not consented to this use of their work, resulting in multiple lawsuits against generative AI developers.In this research, we examine the effect of invoking an artist’s name in the text prompt used to generate an image. We first develop a deep learning model to quantify the impact of including an artist’s name on the perceived aesthetics and preference for an image. We pair this predictive analysis with conjoint analysis to estimate the increase in willingness to pay for products featuring an image that was generated by invoking an artist’s name. Our research offers a means of quantifying the commercial value associated with an artist’s style in products made using generative AI and offers insights relevant to data sourcing to support generative AI.