Large language model (LLM) systems have been shown to stimulate creative thinking among creators, yet empirical research on whether users can seek inspiration in their everyday lives through these technologies is lacking. This paper explores which attributes of LLMs influence inspiration-seeking processes. Focusing on use cases of travel, cooking, and self-care, we interviewed 20 participants as they explored scenarios of these use cases using LLMs. Thematic analysis revealed that the vast data of LLMs inspires users with unexpected ideas, many of which were highly personalized, and inspired participants towards being motivated to act. Participants were also sensitive to the deficiencies of LLMs, and noted how ethical issues associated with these technologies could negatively impact them applying inspirational ideas into practice. We discuss the behavioral patterns of users actively seeking inspiration via LLMs, and provide design opportunities for LLMs that make the inspiration-seeking process more human-centric.
https://dl.acm.org/doi/10.1145/3706598.3713259
Due to the remarkable content generation capabilities, large language models (LLMs) have demonstrated potential in supporting early-stage conceptual design. However, current interaction paradigms often struggle to effectively facilitate multi-round idea exploration and selection, leading to random outputs, unclear iterations, and cognitive overload. To address these challenges, we propose a human-AI co-ideation framework aimed at tracking the evolution of design ideas. This framework leverages a structured idea representation, an analogy-based reasoning mechanism and interactive visualization techniques. It guides both designers and AI to systematically explore design spaces. We also develop a prototype system, IdeationWeb, which integrates an intuitive, mind map-like visual interface and interactive methods to support co-ideation. Our user study validates the framework’s feasibility, demonstrating enhanced collaboration and creativity between humans and AI. Furthermore, we identified collaborative design patterns from user behaviors, providing valuable insights for future human-AI interaction design.
https://dl.acm.org/doi/10.1145/3706598.3713375
A content creator's success depends on understanding their audience, but existing tools fail to provide in-depth insights and actionable feedback necessary for effectively targeting their audience. We present Proxona, an LLM-powered system that transforms static audience comments into interactive, multi-dimensional personas, allowing creators to engage with them to gain insights, gather simulated feedback, and refine content. Proxona distills audience traits from comments, into dimensions (categories) and values (attributes), then clusters them into interactive personas representing audience segments. Technical evaluations show that Proxona generates diverse dimensions and values, enabling the creation of personas that sufficiently reflect the audience and support data grounded conversation. User evaluation with 11 creators confirmed that Proxona helped creators discover hidden audiences, gain persona-informed insights on early-stage content, and allowed them to confidently employ strategies when iteratively creating storylines. Proxona introduces a novel creator-audience interaction framework and fosters a persona-driven, co-creative process.
https://dl.acm.org/doi/10.1145/3706598.3714034
Large shared displays, such as digital whiteboards, are useful for supporting co-located team collaborations by helping members perform cognitive tasks such as brainstorming, organizing ideas, and making comparisons. While recent advancement in Large Language Models (LLMs) has catalyzed AI support for these displays, most existing systems either only offer limited capabilities or diminish human control, neglecting the potential benefits of natural group dynamics. Our formative study identified cognitive challenges teams encounter, such as diverse ideation, knowledge sharing, mutual awareness, idea organization, and synchronization of live discussions with the external workspace. In response, we introduce LADICA, a large shared display interface that helps collaborative teams brainstorm, organize, and analyze ideas through multiple analytical lenses, while fostering mutual awareness of ideas and concepts. Furthermore, LADICA facilitates the real-time extraction of key information from verbal discussions and identifies relevant entities. A lab study confirmed LADICA's usability and usefulness.
https://dl.acm.org/doi/10.1145/3706598.3713289
Involving subject matter experts in prompt engineering can guide LLM outputs toward more helpful, accurate, and tailored content that meets the diverse needs of different domains. However, iterating towards effective prompts can be challenging without adequate interface support for systematic experimentation within specific task contexts. In this work, we introduce PromptHive, a collaborative interface for prompt authoring designed to better connect domain knowledge with prompt engineering through features that encourage rapid iteration on prompt variations. We conducted an evaluation study with ten subject matter experts in math and validated our design through two collaborative prompt writing sessions and a learning gain study with 358 learners. Our results elucidate the prompt iteration process and validate the tool's usability, enabling non-AI experts to craft prompts that generate content comparable to human-authored materials while reducing perceived cognitive load by half and shortening the authoring process from several months to just a few hours.
https://dl.acm.org/doi/10.1145/3706598.3714051
Generative AI in Virtual Reality offers the potential for collaborative object-building, yet challenges remain in aligning AI contributions with user expectations. In particular, users often struggle to understand and collaborate with AI when its actions are not transparently represented. This paper thus explores the co-creative object-building process through a Wizard-of-Oz study, focusing on how AI can effectively convey its intent to users during object customization in Virtual Reality. Inspired by human-to-human collaboration, we focus on three representation modes: the presence of an embodied avatar, whether the AI’s contributions are visualized immediately or incrementally, and whether the areas modified are highlighted in advance. The findings provide insights into how these factors affect user perception and interaction with object-generating AI tools in Virtual Reality as well as satisfaction and ownership of the created objects. The results offer design implications for co-creative world-building systems, aiming to foster more effective and satisfying collaborations between humans and AI in Virtual Reality.
https://dl.acm.org/doi/10.1145/3706598.3713720
Research ideation involves broad exploring and deep refining ideas. Both require deep engagement with literature. Existing tools focus primarily on broad idea generation, yet offer little support for iterative specification, refinement, and evaluation needed to further develop initial ideas. To bridge this gap, we introduce IdeaSynth, a research idea development system that uses LLMs to provide literature-grounded feedback for articulating research problems, solutions, evaluations, and contributions. IdeaSynth represents these idea facets as nodes on a canvas, and allow researchers to iteratively refine them by creating and exploring variations and combinations. Our lab study (𝑁 = 20) showed that participants, while using IdeaSynth, explored more alternative ideas and expanded initial ideas with more details compared to a strong LLM-based baseline. Our deployment study (𝑁 = 7) demonstrated that participants effectively used IdeaSynth for real-world research projects at various ideation stages from developing initial ideas to revising framings of mature manuscripts, highlighting the possibilities to adopt IdeaSynth in researcher’s workflows.
https://dl.acm.org/doi/10.1145/3706598.3714057