AI-Generated Social Media Posts Related to World War I and World War II

This dataset, published in Excel format, contains 307 social media posts collected from Instagram and TikTok. The material was compiled as part of the academic paper “Hypervisual Historical Knowledge: AI-Generated Social Media Content as a Challenge for Digital Public History,” prepared for presentation at the conference Performing Transformation. Shifting Landscapes in Art, Science, and Technology at TU Dresden, 29–30 September 2025. The dataset is structured as a spreadsheet with the following categories of variables: Basic metadata: Post ID, platform, publication date, content type (image/video). Actor and function classifications: Actor type and primary function, following the scheme developed by Mia Berg and Andrea Lorenz within the SocialMediaHistory project. Transparency variables: Indicators of AI labeling by platform or author, disclosure of AI tool used, disclosure of prompt, citation of historical sources, and realism-oriented visualization. Genre variables: Reconstruction of historical figures, animation of historical photographs, colorization, explainer, alternate-history “what if” scenarios, point-of-view perspective, and use of anthropomorphic animals in historical roles. Reference link: Direct URL to the original post. The dataset is designed not to provide statistical exhaustiveness, but rather to constitute a rich and diverse corpus suitable for both qualitative and quantitative analysis. It reflects content discoverable via platform-native search interfaces using combined historical and AI-related keywords, specifically focusing on posts thematically related to World War I and World War II. By making this dataset publicly available via Zenodo, the study seeks to ensure research transparency and to facilitate referencing, verification, and reuse in the broader context of digital public history.