As AI becomes increasingly integrated into how we access information, whether through search engine summaries, large language models, or intelligent interfaces, it’s critical to understand how these tools shape the way we gather, process, and trust information. This project draws on Information Foraging Theory, which models human information-seeking behavior as a process of maximizing reward (useful information) while minimizing cost (time, effort, uncertainty).

We investigate how different interface structures and AI supports impact user strategies and outcomes in information gathering. In the first set of experiments, participants were asked to complete realistic search tasks across three controlled interface types: infinite scroll, paginated layout, and paginated layout with an AI-generated overview. These variations allow us to probe how structural features and AI summarization alter search depth, breadth, and efficiency.

A second experiment examines more naturalistic study behavior. Participants are tasked with learning material to perform well on a melanoma identification quiz, with one of three study conditions: traditional web search, use of a conversational LLM (Gemini), or a hybrid strategy of their choosing. This setup allows us to explore how user autonomy and tool choice affect learning outcomes and search strategies.

Ultimately, this work aims to inform the design of AI-augmented search tools that support effective, trustworthy, and adaptive information foraging, especially in high-stakes domains like health, education, and decision-making.

This work was conducted with Dr. Michael Byrne at the Computer-Human Interaction Lab under Charlie Weeks.