Timelines have traditionally been limited to pre-curated content, restricting users to predefined scopes (e.g., themes or time periods) rather than empowering them to creatively explore evolving areas of curiosity. Leveraging recent advances in AI that can cater to dynamic information needs, we envision more adaptable timelines that provide untethered experiences to enable curiosity-driven exploration. We propose the concept of a generative timeline—a timeline that expands or contracts to adapt to users’ evolving exploration and knowledge-building processes. To instantiate and evaluate this concept, we developed KnowledgeTrail, an interactive system that empowers users to flexibly explore historical events and the process of knowledge formation. A preliminary evaluation with eight participants revealed that KnowledgeTrail supports creative exploration, fosters serendipitous discoveries, and subsequently results in a self-reinforcing cycle of curiosity. This paper contributes the concept of a generative timeline, insights into its design and user experience, and implications for future timeline-based systems that embrace dynamic, user-driven exploration.
