# A Taxonomy of Prompt Modifiers for Text-To-Image Generation

**Authors:** Jonas Oppenlaender  
**arXiv:** [2204.13988](https://arxiv.org/abs/2204.13988)  
**Date:** 2022-04-20  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2204.13988v2/

## Key points

- Text-to-image generation has gained significant attention in recent years
- Deep generative models can synthesize beautiful and intriguing digital images and artworks from textual inputs
- Online communities focused on text-to-image generation and AI-generated art have emerged
- Jonas Oppenlaender presents a novel taxonomy of prompt modifiers used by practitioners in the online community based on a 3-month ethnographic study
- The taxonomy identifies six types of prompt modifiers that provide researchers with a conceptual starting point for investigating the practice of text-to-image generation
- The paper discusses research opportunities presented by this novel creative practice in the field of Human-Computer Interaction (HCI)
- The paper concludes with a discussion of broader implications of prompt engineering from the perspective of Human-AI Interaction (HAI) beyond just text-to-image generation and AI generated art
- The taxonomy provides insight into how prompts can be modified to generate more diverse and creative outputs while highlighting potential ethical concerns related to HAI.

## Summary

The field of text-to-image generation has gained significant attention in recent years, with the ability to synthesize beautiful and intriguing digital images and artworks from textual inputs using deep generative models. As a result, online communities focused on text-to-image generation and AI-generated art have emerged. In this paper titled "A Taxonomy of Prompt Modifiers for Text-To-Image Generation," Jonas Oppenlaender presents a novel taxonomy of prompt modifiers used by practitioners in the online community based on a 3-month ethnographic study. The taxonomy identifies six types of prompt modifiers that provide researchers with a conceptual starting point for investigating the practice of text-to-image generation. The paper also discusses research opportunities presented by this novel creative practice in the field of Human-Computer Interaction (HCI). Furthermore, the paper concludes with a discussion of broader implications of prompt engineering from the perspective of Human-AI Interaction (HAI) beyond just text-to-image generation and AI generated art. The taxonomy provides insight into how prompts can be modified to generate more diverse and creative outputs while highlighting potential ethical concerns related to HAI. Overall, Oppenlaender's work offers valuable insights into the emerging field of text-to image generation and its impact on HAI. It provides researchers and practitioners with a framework for exploring new ways to create compelling visual content using deep generative models while considering ethical implications.