Canvil: Designerly Adaptation for LLM-Powered User Experiences

Advancements in large language models (LLMs) are poised to spark a proliferation of LLM-powered user experiences. In product teams, designers are often tasked with crafting user experiences that align with user needs. To involve designers and leverage their user-centered perspectives to create effective and responsible LLM-powered products, we introduce the practice of designerly adaptation for engaging with LLMs as an adaptable design material. We first identify key characteristics of designerly adaptation through a formative study with designers experienced in designing for LLM-powered products (N= 12). These characteristics are to 1) have a low technical barrier to entry, 2) leverage designers’ unique perspectives bridging users and technology, and 3) encourage model tinkering. Based on this characterization, we build Canvil1, a Figma widget that operationalizes designerly adaptation. Canvil supports structured authoring of system prompts to adapt LLM behavior, testing of adapted models on diverse user inputs, and integration of model outputs into interface designs. We use Canvil as a technology probe in a group-based design study (6 groups, N= 17) to investigate the implications of integrating designerly adaptation into design workflows.

Introduction. A paradigm shift is underway for integrating artificial intelligence (AI) capabilities into everyday user-facing technologies. Large pre-trained AI models, most notably large language models (LLMs), have versatile natural language capabilities that unlock novel interactive techniques and interfaces for more intuitive and customizable user experiences across a wide spectrum of applications [2, 71, 93, 104]. However, these promises also come with numerous concerns. Integrating LLMs into a domain without careful consideration of the user contexts surrounding model use and implementation of behavioral guardrails for the model may result in user experiences that perpetuate societal biases [83], threaten users’ sense of well-being [81, 86], or otherwise do harm [22, 84]. As technology development practitioners, designers2 are uniquely positioned to mitigate these concerns [30, 50, 91, 105, 107].

Discussion / Conclusion. reasoning about interface affordances using research-informed model behavior; these approaches’ promises were amplified once adaptation was embraced as a collaborative practice. Our work illuminates paths for new processes and tools to spotlight designers’ user-centered perspectives and expertise for more responsible and thoughtful deployment of LLM-powered technologies.