Objectives <p>This research note examines the methodological opportunities and challenges of using AI-edited video vignettes in online factorial experiments. Drawing on a randomized experimental study that used a web-based video vignette survey to test the causal effects of police officer race and behavior on public perceptions and cooperation, we highlight lessons learned and considerations for future research using similar designs.</p> Methods <p>The experimental (parent) study employed factorial video vignettes embedded in a Qualtrics survey, with randomization features, and recruited participants via the Connect platform. Key methodological elements, including video production, randomization, sampling, and ethical safeguards, are reviewed alongside implementation challenges in this research note.</p> Results <p>This study achieved high completion rates and low bounce rates, with a demographically diverse sample. Participants rated the survey positively across user experience, clarity, and fairness of compensation. Methodological strengths included mainly the innovative use of AI-modified videos for experimental stimuli, quota sampling, and quality control. Challenges were primarily technical, as current AI tools remain limited in their ability to convincingly alter officer race in video.</p> Conclusions <p>Though AI-based racial modifications in video remain challenging, factorial video vignette designs in online experiments offer significant promise for causal inference.</p>