Exploring User Perceptions of Generative AI in Data Storytelling

Data storytelling combines narrative techniques that engage the audience with data visualization that provides meaningful insights, aiming to provide insights in an actionable way. Like various disciplines in the last few years, Generative AI use cases are being explored within data storytelling to some extent with limited success. This study investigates whether LLMs can be used by people in this creative of crafting data-driven stories that aim to assist in decision-making. To evaluate this, we developed a three-stage survey that captures participants' initial opinions on using Generative AI tools, then subject them to an intervention where they create a data story using ChatGPT and finally conclude with a post-survey that accumulates their thoughts on Generative AI and storytelling after the intervention. Our research examines how prior experience with data visualization and Generative AI shapes people's attitudes toward using these tools and whether hands-on experience affects their perceptions.The findings suggest that individuals with previous experience in Generative AI are inclined to adopt these tools for data storytelling. However, those who have used LLMs specifically for visualizations might not adopt them as readily. We also discovered that users find responses from tools like ChatGPT most helpful in the initial stages of data storytelling. Further, if they perceive the results as demonstrating expertise, they are more likely to embrace these Generative AI.