ChatbotLab: An Open Framework for Human-Like Chatbot Interaction and Survey Data Collection

<p>Despite growing interest in studying human-AI interaction and conducting conversational interventions at scale, technical barriers in deploying and managing large language model (LLM) systems limit their use in behavioral research. In this tutorial, we present ChatbotLab, an open-source software package that provides an end-to-end framework for conducting chatbot-based research without requiring programming expertise. ChatbotLab embeds LLM-powered conversations directly into standard crowdsourcing services (Prolific, MTurk) and survey platforms (Qualtrics, REDCap, LimeSurvey), handling the technical infrastructure that custom solutions often overlook: secure data linkage between surveys and conversations, reproducible study records that capture each conversation's full bot configuration, content moderation, standardized data export for downstream analysis, and server load testing for managing thousands of simultaneous conversations. We guide researchers through the complete workflow, from deploying a chatbot server and configuring naturalistic conversation features (typing delays, message chunking, and personas) to integrating conversations into existing study pipelines and exporting data for analysis. By lowering the technical barriers to chatbot-based research, ChatbotLab allows researchers to focus on their research questions rather than infrastructure.</p>