By Wyn Brodersen and Maik Fielitz
Antisemitic content on TikTok rarely announces itself. It arrives as a number, an emoji, a joke about a hollow earth — and that is exactly what makes it spread.
In February 2026, a local politician for Germany’s far-right AfD shared a video on Instagram. Its text asked whether “an Austrian painter” had been right that elites drink the blood of children, with a 🧃 emoji in connection with the word “elite.”
Three references, none of them explicit. The painter can be seen as a nod to Adolf Hitler. The juice box works as a code for Jews. And the blood drinking elites is a known reference to antisemitic conspiracy theories rehashed by the QAnon cult nowadays. Confronted with national outrage, the politician said she had not recognised any of it — she hadn’t known Hitler painted, and what this emoji would stand for.
Whether or not this is true, the defence is instructive. Coded antisemitism produces exactly this deadlock: the meaning is unmistakable to those who already know the references, and deniable to everyone else. The results of our study on antisemitism on TikTok ask how that deadlock is manufactured at scale on the platform where it works best.
Antisemitism on TikTok: A Challenging Object of Study
TikTok is difficult to research and difficult to moderate for the same reason. Its algorithm leans on observed behaviour rather than social graphs, so reach depends less on who you are than on whether your format latches onto something already trending. Extremist material adapted to a popular sound, meme template, or in-joke can therefore reach audiences that no extremist account could otherwise touch — and audiences interact with it without necessarily consenting to the respective worldview.
Meaning on TikTok is also assembled from parts. Image, caption, audio, hashtags and on-screen text each carry a fragment; the antisemitic reading emerges only from the combination. Automated moderation and keyword-based research both tend to miss what is distributed across modalities. This also made it difficult to filter antisemitic content. Working with computational methods, we decided to take hashtags as gateways.
We collected 20,965 TikTok posts published between January 2025 and February 2026, retrieved through 34 hashtags selected from prior literature and material scoping. We kept German- and English-language posts with more than 500 views from accounts registered in North America, Europe, North Africa and the Middle East. A weighted random sample of 2,446 posts was then coded by hand.
One caveat matters throughout: hashtags are a keyhole, not a window. They capture only the content their authors chose to tag, and our sample is deliberately non-representative. Nothing here estimates how common antisemitism is on TikTok overall.
Trends as a Lever for Amplification
Hashtags follow temporal patterns. For example, some increases can be seen in the summer of 2025 for the hashtags ‘Zionism’, ‘Deepstate’, and ‘Israhell’. These increases may be connected to the war between Israel and Iran, which escalated in June 2025.
The most prominent hashtag in our data — dominant since autumn 2025 — was #Agartha, a story about a mythical Aryan civilisation supposedly inhabiting a hollow earth, reachable through portals in the Himalayas or Antarctica. The trend braids together theosophical motifs, occult “Vril” energy borrowed from an 1871 science-fiction novel, and the Nazi esotericism that recycled both. Videos feature AI-generated maps of Hyperborea, SS uniforms and Nazi flying saucers, usually played for laughs.
The irony is not a disguise of the ideology, but rather the engine. The pleasure lies in generating ever more absurd variations whose humour depends on being unintelligible to outsiders. While the Aryan ideal can itself be interpreted as antisemitic, overt antisemitism is not apparent in every post. Some posts instead rely on more subtle references. For instance, one road sign displays the distances to Auschwitz, Valhalla, and Agartha alongside the number 271. Other posts feature the White Monster energy drink, one of the most common visual motifs within the Agartha trend, and likewise associate it with the number 271. The number derives from a historical document compiling death certificates from several, but not all, concentration camps. In these posts, it is used to downplay the Holocaust by falsely presenting 271,000 as the total number of Jewish victims rather than approximately six million.
Ambiguity Outperforms Explicitness
Only about 3% of the coded posts were unambiguously antisemitic — content degrading or stereotyping “Jews because they are Jews”. A further 19% fell into a second category, entitled ambiguously antisemitic, that included stereotyping allusions, phonetic games, numeric codes, and communicative detours that allowed for an antisemitic interpretation without being explicitly antisemitic.
Figure 1: Classification of the posts according to relevance and potential antisemitic references.
While explicitly antisemitic posts clustered around a median of some 1,800 views, with the middle half topping out near 16,500, ambiguous posts spread across a much wider range, with the middle half extending beyond 27,000.
Figure 2: Reach of explicit and ambiguous antisemitic posts (logarithmic scale).
However, several high-reach outliers have been viewed millions of times. These include another road sign and an Instagram search interface, both of which link Agartha and Israeli institutions to the code 271. The first received more than 2.6 million views, and the second exceeded 5.6 million, with each attracting several hundred thousand likes and thousands of comments.
Ambiguous posts were also more likely than explicit ones to minimise the Holocaust or nod approvingly at Nazi exterminationist antisemitism, suggesting that humor enables transgression. Overall, conspiracy narratives ran at roughly 44% in both categories.
Figure 3: Forms of manifestation in explicit and ambiguous antisemitic posts.
Speaking Up Against Online Antisemitism
We also found pushback: 9% of relevant posts reclaimed the same hashtags for Jewish life, education about the codes, or satire deflating the fantasy of Jewish omnipotence. The posts had different goals and used various strategies. Some relied on an element of surprise, making them not immediately recognisable as interventions. Drawing on antisemitic motifs and modes of representation, they aimed to expose the underlying mechanisms and strategies.
Others parodied antisemitic stereotypes by pushing them to absurd extremes or responded directly to antisemitic content through video-stitches. In addition, some posts used these hashtags to present Jewish life in a positive light, seeking to demonstrate their own agency beyond the perspective of victimisation.
Research also shows, however, that these strategies face the challenge of TikTok’s moderation not always adequately accounting for context. As a result, antisemitic content may remain online while interventionist or educational posts are removed.
Hidden in Plain Sight
None of the highest-reaching posts in our dataset came from known extremist actors. Antisemitic motifs did not spread as a programme; they spread by attaching themselves to whatever was already moving — gym content, looksmaxxing, gaming clips, AI slop, an eighties pop remix.
Historians of nineteenth-century Europe called this antisemitism a cultural code: a signal of belonging to an anti-liberal, anti-modern camp, in which hostility to Jews does work for the in-group more than it describes any actual target. The digital version looks much the same, minus the camp. What holds these communities together is not a political position but a shared fluency in references.
That has an uncomfortable implication for prevention. Debunking specific claims addresses the content while leaving intact what actually carries it: the play, the belonging, the aesthetic pull. Understanding online antisemitism now requires literacy in digital culture as much as in ideology — and the fact that it is merely one trend among many is precisely the measure of how normal it has become.
The full study is available here: Wyn Brodersen, Michael Schmidt & Maik Fielitz, “Portionierter Judenhass. Eine Vermessung des Antisemitismus auf TikTok,” in Machine Against the Rage No. 9, Spring 2026. DOI: 10.58668/matr/09.2.
Wyn Brodersen is a sociologist and researcher at the Jena Institute for Democracy and Civil Society. His research focuses on the digital dimensions of radicalization processes, far-right ideology and communication, as well as disinformation. He is part of the editorial team of the online magazine Machine Against the Rage.
Maik Fielitz is a social scientist and conflict researcher. He is the head of the research unit on digital conflict studies at the Jena Institute for Democracy and Civil Society, as well as co-editor of the online magazine Machine Against the Rage. His research examines the ways in which digital technologies and digital cultures influence the emergence and evolution of right-wing extremism, as well as the strategies employed by liberal democracies to counter authoritarian tendencies in online contexts. Together with Wyn Brodersen, he co-authored the Handbook on Online Antisemitism.