Daejeon, July 27–31
In early spring of 2025, a viral history POV TikTok trend emerged. The videos have titles such as „POV: You wake up in Pompeii on Eruption Day“ or „POV: You wake up as Queen Cleopatra on her last day“ and consist of a sequence of point-of-view shot scenes, illustrating these historical scenarios. They are a manifestation of so-called “AI slop”, “low-quality digital content produced, usually in quantity, by means of AI” (Baltes et al. 2026: 1).
Already since the controversial #HolocaustChallenge, during which creators staged themselves as Holocaust victims in the summer of 2020 (Divon / Ebbrecht-Hartmann 2023), a scholarly and public awareness of history-related content on TikTok has begun to develop (Berg / Lorenz 2025: 181). TikTok has become an increasingly popular subject of research. However, studies on specific topics and case studies are rare (Cervi et al. 2023: 204). It is difficult to keep up with TikTok’s fast-paced nature, and researchers encounter technical, legal, and ethical hurdles. Although TikTok provides a research APIhttps://developers.tiktok.com/products/research-api/ [08.12.25], access is limited, and content archiving is not supported (Berg / Lorenz 2025: 182); neither is it possible to obtain the videos themselves. Thus, studies typically rely on the method of “digital ethnography”: a new account is created, content is consumed via access to the ForYouPage, hashtags, or accounts within a set period of time, and documented as “field observations” (e.g., Berg/Lorenz 2025, Ackermann 2025). Research using the API is less common, even though it offers more comprehensive access to the content. For instance, Jan-Robbert Adriaansen (2022) provides an overview of history-related content using the TikTok research API. This proposal seeks to advance research on history-related content on TikTok and to engage with ongoing debates in public history, especially regarding authenticity, history education, and historical representation. They are currently being conducted not only in relation to AI and social media, but also regarding virtual reality (see most recently Günther 2025).
As part of a preliminary study, outlined in Brolich / Neovesky (2026), we have compiled an English-language dataset encompassing 5,565 history POV videos, using the TikTok API. A frequency analysis of the captions revealed the most commonly chosen topics, as illustrated in Fig. 1. It suggests that creators primarily focus on generally popular historical and emotionally charged themes, such as disasters and wars, reflecting a tendency towards sensationalism, potentially as a strategy to increase engagement and virality (Neubert 2024: 141f.).
This is not particularly surprising, as over 75 % of the videos were uploaded in the US, the UK, Germany, or France alone, highlighting the benefits of expanding the dataset using a more nuanced, multi-language approach.
For this contribution, we extend our previous method through a multilingual data collection and analysis framework using the TikTok Research API, focusing on content from several regions – namely the US, South America (Brazil, Argentina, Colombia, Chile, Peru, Venezuela, Ecuador, Bolivia, Paraguay, and Uruguay), German-speaking Europe (Germany, Austria, Switzerland), South Korea, and Indonesia. By incorporating open source translation models from the No Language Left Behindhttps://ai.meta.com/research/no-language-left-behind/ [06.05.26] project and critically assessing their impact, we build a nuanced history POV dataset. It allows us to obtain a more comprehensive understanding of the trend and its topics through frequency and spatiotemporal analysis. We can also examine how certain topics are related to engagement metrics: are some topics more successful than others?
Additionally, by comparing our previous English-only method and the new multi-language approach, we can investigate new research questions: How did the trend unfold: were videos merely translated from English or did creators produce their own original content? Who contributes to a trend like this? How are the language and provenance of the videos related to their thematic focuses? Is there a difference in the choice of topics by creators from the area instead of just talking about the area?
The results will be situated within broader debates on the “generated past,” showing how the History-POV trend on TikTok exemplifies the influence of widely accessible AI technologies on the practice of doing history in social media, where the past is increasingly conceived as a world that can be algorithmically generated, affectively activated, and performatively enacted for cultural styling (Matei 2024: 14).