Donghoon Shin, Alice Gao, Rock Yuren Pang, Jaewook Lee, Katharina Reinecke, Emily Tseng
4 min
Abstract
Generative AI is known for its tendency to homogenize, often reproducing dominant style conventions found in training data. However, it remains unclear how these homogenizing effects extend to complex structural tasks like web design. As lay creators increasingly turn to LLMs to 'vibe-code' websites -- prompting for aesthetic and functional goals rather than writing code -- they may inadvertently narrow the diversity of their designs, and limit creative expression throughout the internet. In this paper, we interrogate the possibility of design homogenization in web vibe coding. We first characterize the vibe coding lifecycle, pinpointing stages where homogenization risks may arise. We then conduct a sociotechnical risk analysis unpacking the potential harms of web vibe coding and their interaction with design homogenization. We identify that the push for frictionless generation can exacerbate homogenization and its harms. Finally, we propose a mitigation framework centered on the idea of productive friction. Through case studies at the micro, meso, and macro levels, we show how centering productive friction can empower creators to challenge default outputs and preserve diverse expression in AI-mediated web design.
Alex: So risks at key steps trap users in generic designs?
Sam: Yes. Representational harms make diverse styles seem wrong. Quality glitches push back to reliable defaults. Cognitive harms frustrate users into grabbing the simple option. Loss of agency locks in the first version when code gets complex.
Alex: How do they suggest fixes at those steps?
Sam: They propose productive friction—pauses that prompt thought without losing speed. Think of it like a GPS suggesting a detour only after checking your goal. For one creator, the tool asks clarifying questions or shows style options, like mood boards of dense versus sparse layouts. For teams, it compares to a brand guide.
Alex: And for the wider web?
Sam: At the largest scale, sameness could fill the internet, leading to model collapse where future AIs train on less varied data—like a library of identical books. Fixes include picking style anchors upfront, like a color palette, and tagging outputs.
Alex: How solid is the evidence?
Sam: It's from user discussions and tool tests—a strong conceptual map. But it lacks user studies on these frictions, so more empirical work could test them.
Alex: This paper maps how easy AI tools risk narrowing web designs but points to targeted pauses for more variety. A clear step for better tools. Thanks, Sam—and thanks for listening to ResearchPod.