How to Create Stunning Images with ChatGPT

Why Read This from an HCI Perspective

This article helps readers see AI not just as a ‘tool that can draw well,’ but as an interaction in which users specify, revise, and interpret the results. For HCI/UX practitioners and researchers, it’s important that prompt writing, iterative refinement, and result evaluation form a single user journey. In particular, with generative AI, the quality of the experience depends not only on model performance, but also on when and how users intervene—and where they feel a sense of control.

CIT's Commentary

The core of this article is less about the quality of AI-generated images and more about how users adjust and learn from the outputs. In other words, it’s not only about ‘getting good results,’ but also about having a ‘good revision process.’ The flow of editing prompts and trying again isn’t just a feature—it’s a feedback loop. If the interface isn’t user-friendly, people may end up relying on luck without understanding the results. Conversely, when the interface clearly shows the reasons for changes and what to modify, users’ sense of control and trust increase. As generative AI becomes more widely used, the real competitive advantage of a product will be how much it can reduce the fatigue of these iterative tasks. Especially in Korea’s service environment—where rapid exploration and mobile-first usage are common—design that communicates intent well with short inputs and makes it easy to compare results becomes even more important.

Questions to Consider While Reading

• Q.What is the most effective interface for helping users understand why a certain image result came out the way it did during the generation process?
• Q.What kind of feedback structure is needed to reduce the fatigue of repeatedly revising prompts while still maintaining users’ sense of control?
• Q.In Korea’s mobile service environment, how should the interaction model of generative AI image tools differ from that of global products?