The Future of Ad Research with Generative AI: MADE Framework Made Simple

Artificial intelligence (AI), particularly generative AI as applied in content creation, has become a research topic of significant interest. While discussions about its quality, integrity, and copyright issues in research have arisen, the potential of generative AI to assist researchers in developing experimental stimuli is often overlooked. Zeph et al. (2024) introduced the MADE (Mapping, Assembling, Demonstrating, Executing) framework, aiming to provide a comprehensive set of best practices for the ethical and effective use of generative AI in creating experimental stimuli for advertising research.

The Realism–Control Paradox in Experimental Advertising Stimulus Design

Experimental advertising research has long faced the paradox between realism and control. To ensure the scientific rigor of research, experimental designs often require strict control over stimulus materials, manipulating only the target variables. However, if experimental advertising stimuli differ significantly from real-world advertisements, their ecological validity may be questioned, thereby limiting the practical value of the research findings. Geuens and De Pelsmacker (2017) pointed out that high-quality and realistic experimental advertising stimuli are crucial for advertising research, but academically produced advertisements often fall short of this standard.

However, Generative AI, with its ability to generate high-quality, research-relevant text, images, and videos, offers new opportunities to address the aforementioned realism challenge. Advanced algorithms such as generative adversarial networks (GANs) and Transformer models (like GPT) can rapidly generate large amounts of content. By leveraging generative AI, researchers can more easily manipulate specific elements in advertisements, such as changing the age or expression of a model, or adjusting the type or color of a product, while maintaining consistency in other aspects of the advertisement. This capability is expected to enhance the construct validity, internal validity, external validity, and ecological validity of experimental advertising research.

AI-assisted experimental stimuli design

Enhancing Experimental Design with Generative AI: The MADE Framework for Stimulus Development

Empirical studies have shown that the MADE framework can assist researchers in using AI to design advertisements that perform comparably to real advertisements across multiple dimensions of ecological validity—including quality, appropriateness, and realism. This research provides preliminary support for the potential of generative AI to enhance ecological validity in experimental advertising research. To support responsible and effective use, the MADE framework offers a practical process that balances creative flexibility with academic rigor.

MADE consists of four key phases:

•     Mapping: Define the variables you want to manipulate and consider the ad’s context—target audience, platform, and cultural setting. Optionally, you can reference a real ad to guide your AI prompt.

•     Assembling: Choose the right AI tool (text, image, or video), craft detailed prompts, evaluate the generated output, and manually refine it if needed—for example, by adding logos or adjusting the visuals.

•     Demonstrating: Pretest your stimuli to check clarity, realism, and potential confounds. Adjust based on feedback before moving to the full study.

•     Executing: Use the stimuli in your main experiment and ensure all ethical considerations are addressed, including disclosure of AI use, copyright issues, and data privacy.

• 

  MADE Framework

  Using Generative AI in Advertising Experiments? Try the MADE Framework

  The MADE framework offers a structured guide for developing experimental materials in advertising research, enabling researchers to use generative AI more efficiently, responsibly, and systematically. While AI-generated content may still require time and manual refinement, its low cost and relatively easy learning curve make it a promising tool for improving the overall quality of experimental advertising studies.

  Despite ongoing methodological challenges, the application of generative AI presents significant opportunities to enhance the quality and realism of experimental stimuli—marking a new chapter in addressing the “realism–control paradox.” As the field of advertising research continues to adapt to these technological changes, ongoing exploration will be essential to optimize the use of AI tools in creating effective experimental materials.

  References：

  Geuens, M., & De Pelsmacker, P. (2017). Planning and Conducting Experimental Advertising Research and Questionnaire Design. Journal of Advertising, 46(1), 83–100. https://doi.org/10.1080/00913367.2016.1225233

  van Berlo, Z. M. C., Campbell, C., & Voorveld, H. A. M. (2024). The MADE Framework: Best Practices for Creating Effective Experimental Stimuli Using Generative AI. Journal of Advertising, 53(5), 732–753. https://doi.org/10.1080/00913367.2024.2397777