Anyone who has watched the news over the past few years has heard that antisemitism is rising, but the harder question is which specific narratives are gaining traction and which are receding. We tracked US search interest in 18 terms that can function as inputs to antisemitism, from conspiracy proper-nouns to contested framings of Israel, and measured how much each moved over the past year. The answer: 15 of the 18 drew more search interest than the year before, a rise that is broad rather than confined to one theme, with the “great replacement” conspiracy leading at about 240% year over year and “anti-Zionism” close behind at about 209%.
This is the first reading of the Institute’s search-attention tally, a running record of US Google search interest in these terms. It measures how much each narrative is being searched, a signal of salience, not a measure of antisemitic belief, and the relationship is correlational.
Key Findings
- Of the 18 tracked narratives, 15 drew more search interest over the last year than the year before; the rise is broad rather than confined to one theme (Hanover Institute search-attention tally, Google Trends, June 2026).
- The single steepest climber is the “great replacement” conspiracy, up about 240% year over year, a narrative independent research has tied to offline hate crime (Aziani and colleagues, 2025; Verma and colleagues, 2024).
- Israel-framed terms rose sharply around the war: “anti-Zionism” up about 209% and “israel genocide” up about 126%, both jumping after October 2023.
- Coded classical tropes climbed too: “jewish power” up about 147%, “holocaust denial” about 117%, “khazars” about 95%, and “blood libel” about 70%.
- A few framings cooled: searches for “george soros” fell about 27% and “israel apartheid” about 8%, even as the broader field rose.
What is this tally, and what does it measure?
It is a running record of US Google search interest, pulled monthly back to 2015, for a curated list of terms that can function as inputs to antisemitism: conspiracy proper-nouns (the “great replacement,” Soros, the Rothschilds), coded classical tropes (blood libel, the Khazar myth, “jewish power”), and contested framings of Israel (anti-Zionism, the genocide and apartheid framings). The lane filter is the input lens used across the Institute’s work: does the term plausibly fuel antisemitism? The full method, including the year-over-year change used here, is set out in the methodology section.
One thing the tally is not is a measure of antisemitism itself. A rising search for “blood libel” may be a student looking it up, a journalist, or someone spreading it; search interest measures attention and salience, not endorsement. Read it as a leading indicator of which narratives are active in the discourse, one signal to be set beside the measured-content and incident evidence, never as a count of antisemitic belief.
Which narratives are rising fastest?
The climb is led by the conspiracy and contested-framing terms. Year-over-year, search interest in the “great replacement” rose about 240%, the most of any tracked term, followed by anti-Zionism (about 209%), “jewish power” (about 147%), the genocide framing of Israel (about 126%), and holocaust denial (about 117%). The coded classical tropes are not far behind, with the Khazar myth up about 95% and blood libel about 70%.
The three steepest climbers surged at different times, which is itself informative. The Israel-framed terms jumped almost vertically after October 2023 and have stayed high, the pattern of an event-driven spike. The “great replacement” has a different shape: a brief spike in 2019, the year of the Christchurch attack whose perpetrator titled his manifesto with the phrase, then a quieter period, and then a steady organic climb through 2025 and 2026 to its highest level on record.
Why the “great replacement” climb matters most
Of the rising terms, the “great replacement” is the one with the strongest measured link to offline harm, which is why its climb is the headline. In an analysis of online search demand for 36 racial and political conspiracy theories against weekly hate-crime counts, the Great Replacement theory produced the single largest improvement in predicting registered hate crimes, reducing prediction error by up to 6.42% at a three-week lag (Aziani and colleagues, 2025, a preprint). And in a study of US presidential-election windows, the new antisemitic content that surged aligned with Great Replacement narratives attributing demographic change to Jewish influence (Verma and colleagues, 2024, peer-reviewed). A rising search signal for this particular narrative is not just chatter; it tracks a documented driver.
Which framings are cooling?
Not everything rose. Search interest in “george soros” fell about 27% year over year, and the apartheid framing of Israel fell about 8%, even as the genocide framing of Israel rose. The Soros decline is the clearest: searches for the financier, long a focal point of the hidden-power conspiracy, are now running far below their late-2010s peak. Whether attention has shifted from the named figure to the broader “globalists” framing, which was roughly flat, is a question the tally can track going forward but cannot answer from one reading.
What this does and does not show
The tally shows where attention is moving, and it converges with other measures the Institute tracks: the secular rise in attention to antisemitism overall and the way antisemitism spikes around conflict events. What it does not show is who is searching or why. A rising search line is a rising-salience signal, consistent with a narrative spreading, but it cannot distinguish a believer from a researcher, and it is not a count of antisemitic incidents or attitudes. It is one instrument on the dashboard, read alongside the content-analysis and incident evidence, never on its own.
Methodology
This report draws on a first-party dataset, the Institute’s search-attention tally. Provenance: US Google search interest retrieved through the Google Trends interface (web search, United States), pulled monthly from January 2015, for 18 curated terms selected under the input lens (does the term plausibly fuel antisemitism?). The tally is a running record: each pull is persisted, so the series extends over time.
The metric reported here is year-over-year change, the average indexed interest over the most recent 12 months against the average over the prior 12 months. Limitations: Google Trends values are indexed from 0 to 100 against each term’s own peak in the window, so they are comparable within a term over time, which is the signal used here, but not across terms in absolute size. A very large percentage can reflect a rise from a low base, and search interest measures attention, not antisemitic belief or behavior. The relationship between rising search salience and antisemitism is correlational; this report makes no causal claim.
Conclusion
So which narratives are gaining traction and which are receding? The tally answers plainly: the rise is broad, not confined to one theme. 15 of the 18 tracked terms drew more US search interest over the past year than the year before, and the climb is led by the terms a reader would least want to see climbing.
That breadth holds across every axis the tally measured. The conspiracy and contested-framing terms top the table, the “great replacement” up about 240% and “anti-Zionism” about 209%, with the genocide framing of Israel up about 126%, and the coded classical tropes are not far behind, “jewish power” about 147%, holocaust denial about 117%, the Khazar myth about 95%, and blood libel about 70%. The three steepest climbers moved on different clocks: the Israel-framed terms jumped after October 2023 and stayed high, while the “great replacement” climbed steadily through 2025 and 2026 to its highest level on record. Not everything rose, “george soros” fell about 27% and the apartheid framing about 8%, but the field as a whole moved up, and the term leading it is the one with the strongest measured tie to offline harm, reducing hate-crime prediction error by up to 6.42% at a three-week lag (Aziani and colleagues, 2025, a preprint).
These are not obscure phrases. They are the conspiracy proper-nouns and coded tropes that live well outside any search box, in the everyday joke about who controls what, in the aside about hidden power, in the replacement story repeated for a laugh or in earnest. Search interest is one of the inputs that marks where a narrative is active before it surfaces as coded content, which is why the Institute tracks it. The tally fixes the size of the shift and the shape of it: 15 of 18 narratives up, the steepest climb on the term with the strongest measured tie to offline hate crime. What it cannot fix is what sits underneath the climbing line. When a society looks up the replacement story 240% more than it did a year earlier, and looks hardest at the narrative that best tracks where hate crime lands, is it the belief that is spreading, or only the wish to understand a thing already in the air?
Frequently Asked Questions
Does rising search interest mean antisemitism is increasing?
Not by itself. Search interest measures how much a term is being looked up, which signals salience, not endorsement; a person searching “blood libel” may be researching it. It is a leading indicator to be read beside content-analysis and incident data, several of which are also rising, not a direct count of antisemitism.
Which narrative is rising fastest?
The “great replacement” conspiracy, up about 240% in US search interest year over year, the most of any term the Institute tracks. It is also the rising term with the strongest measured link to offline hate crime (Aziani and colleagues, 2025).
Why did searches for “george soros” fall?
Search interest in the financier fell about 27% year over year, running well below its late-2010s peak, even as the broader field of antisemitism-driving terms rose. The tally cannot say why from one reading; it can track whether attention has shifted to the broader “globalists” framing over time.
Is this the same as measuring antisemitism online?
No. Measuring antisemitism online means coding content for antisemitic meaning, which the Institute’s other reports cover. This tally measures search attention to a set of narratives, a different and complementary signal that shows where interest is moving before it shows up as coded content.
Sources
- Aziani, Lo Giudice, Yazdi, 2025. Conspiracy to Commit: Information Pollution, Artificial Intelligence, and Real-World Hate Crime. arXiv:2507.07884. Preprint.
- Hanover Institute, 2026. Search-attention tally. US Google search interest for 18 antisemitism-driving terms, retrieved June 2026 via Google Trends. First-party dataset.
- Verma, Sear, Johnson, 2024. How U.S. Presidential Elections Strengthen Global Hate Networks. npj Complexity, DOI 10.1038/s44260-024-00018-8. Peer-reviewed.