Generative Experiments: A Novel Approach to Studying Consumer-AI Interactions
(A2025-126454)

AUTHORS

Siv Pedersen, Copenhagen Business School; Constant Pieters, Copenhagen Business School; Felix Eggers, Copenhagen Business School

ABSTRACT

As Generative AI (GenAI) becomes increasingly accessible to consumers, understanding its implications for marketing is essential. This paper introduces Generative Experiments, a novel methodological approach for studying consumer-AI interactions under realistic yet controlled conditions. Departing from vignette and field experiments, Generative Experiments involve configuring GenAI systems with tailored instructions—potentially involving experimental manipulations— to simulate and measure interactions. In practice, this approach uses a custom GenAI-integrated survey interface to capture both behavioral data, such as consumer prompts, and self-reported data. We demonstrate that Generative Experiments can effectively manipulate and assess distinct interaction styles, such as warmth and competence, while limiting threats to validity. By offering a user-friendly and versatile tool, this work enhances the marketing research toolkit to enable more accurate insights into consumer-AI dynamics.