Beyond magic: Prompting for style as affordance actualization in visual generative media
- For policymakers
- Summary created: 2025
This research studied how people who generate AI images add style modifiers to their prompts and examined what this reveals about how people interact with generative AI.
.jpg)
This study examined how users of the visual generative AI system Midjourney “prompt for style” to achieve desired aesthetic outcomes, through user discussions on the Midjourney’s Discord server. I did this to explore the dynamics of human-model interaction and discuss the implications of user prompting practices for style recontextualization without proper attribution.
Insights
Style modifiers function as an entry point into understanding human-model interaction, positioning AI image generation as a complex process of affordance actualization rather than simply magical or deterministic.
Evidence
This means that using style modifiers is not just a technical command but a key part of the interaction between the user and the AI system, where the user’s goals shape the final output. AI image generation is presented as a process of realizing possibilities, rather than a simple, predictable result.
What it means
The paper frames style modifiers as an affordance that users actualize to describe how people perceive and utilize the technological possibilities offered by the AI system. A human prompter acts as a goal-oriented actor, perceiving a style modifier as an opportunity to be acted upon to achieve a specific stylistic outcome.
While visual generative media offers potential for expanding creative expression, the practice of prompting for style often results in generating visual aesthetics that mimics existing cultural artifacts.
Evidence
Style modifiers allow users to mimic the style of human artists, serving as shortcuts to achieving desired visual outputs through associations made from training data containing copyrighted artworks scraped from the web. For example, users can reference names of specific artists to reproduce characteristic styles.
What it means
This means that AI image generation operates through “procedural imitation” of existing works, where the model looks for statistical patterns from its training data. Consequently, what appears to be a creative output is actually a “form of aesthetic mimicry.”
Proposed action
AI image generation platforms must implement clear transparency systems that automatically identify and credit the artistic styles or cultural works being replicated in generated outputs.
There should be more focus on the development of alternative prompting frameworks that encourage creative exploration rather than simple replication of the styles of the living artists.
Share your thoughts
You must be logged in to ask a question. Make an account.
Are you a researcher looking to make a real-world impact? Join Acume and transform your research into a practical summary.