Exploring the Impact of Interface Design on User Experience in AI-Assisted Image Generation: An Empirical Study

The rapid advancement of AI-assisted image generation tools has democratized visual creation, yet user experience (UX) challenges persist due to suboptimal interface design. This study examines how visual feedback types (Gradual Reveal, Linear Progress Indicator, Animated Feedback) and control layouts (Linear vs. Mapped) influence task performance, subjective experience, and visual attention in a simulated AI image generation environment. Using a 3 × 2 mixed factorial design with 60 participants, we measured task completion time, error rate, usability (SUS), Smoothness, and Enjoyment, with supplementary eye-tracking data from 20 participants to explore visual attention patterns. Results showed that Mapped Layout significantly reduced task time and errors while improving usability. Animated Feedback enhanced perceived Smoothness and Enjoyment, with interaction effects suggesting layout–feedback synergies. Eye-tracking patterns were consistent with behavioral and subjective findings. These results offer design guidance for human-centered generative AI interfaces, highlighting the importance of spatial mapping and engaging feedback.