This study explores generative image AI’s capacity to rapidly create multiple realistic renderings for human-robot interaction (HRI) research. By utilizing freehand sketches and employing an iterative process of rendering generation and refinements in Vizcom and Photoshop, variations of service robot designs for four different use contexts were created and visualized as user evaluation media to understand the perception of robot morphology. The AI-assisted workflow, from sketching to rendering, is examined by investigating each step with outcomes, followed by reflections on the strengths and challenges of co-creating with generative image AI. This study contributes to design research and practice by proposing a novel application of generative image AI for creating variations of situated robot representations in HRI and by offering insights and suggestions for its effective use.