Antisemitism on social media has never been higher. Three years after Hamas terrorists murdered some 1,200 people in southern Israel, the Online Hate Prevention Institute (OHPI), an Australian charity that has tracked online Jew-hatred since 2022, has found that none of the platforms it monitored before the attack have returned to their earlier levels. The most recent three months were the worst it has recorded.
July through September 2026 produced the highest levels of antisemitic content in any three-month period since OHPI began monitoring. The report, titled “Analysing the new normal,” is the institute’s fifth in the series and was written by OHPI CEO Dr. Andre Oboler, analyst Jasmine Beinart and Eliyahou Roth of the Online Hate Task Force. It draws on more than 29,500 items of antisemitic content collected from 11 platforms: Facebook, Instagram, X, TikTok, YouTube, LinkedIn, Reddit, Telegram, Gab, BitChute and, for the first time, Threads.
OHPI’s analysts spend eight hours a month on each platform, starting with searches for categorically antisemitic terms and then following the users who engaged with that content. Every item is classified using a taxonomy drawn from the International Holocaust Remembrance Alliance (IHRA) Working Definition of Antisemitism, now Australia’s official definition.
“The most significant finding is that antisemitism is up, not down,” Oboler told J-Wire.
A rising floor
Antisemitism spiked in the months after October 7, dipped slightly through the rest of 2024, then climbed again. By August 2025, it had passed the levels seen immediately after the massacre. OHPI found that each decline left the baseline a little higher than before, and over three years the overall trend has been upward. The institute attributes this to both rising antisemitism and declining efforts by the platforms to fight it.
The trajectory differs by platform. Instagram, TikTok and YouTube have kept getting worse since the initial surge. Facebook, X, Gab, Telegram, Reddit and BitChute have stayed at their elevated post-October 7 levels.
The old hatred dominates
The report’s most pointed finding concerns what kind of antisemitism is flooding social media. Critics of antisemitism monitoring often argue that the post-October 7 numbers are inflated by counting legitimate criticism of Israel as Jew-hatred. OHPI says its data shows the opposite. Traditional antisemitism, including Jewish power conspiracies, blood libels, and the accusation that Jews killed Jesus, is the dominant category on almost every platform and in total. It also rose more after October 7 than any other category. According to The Jewish Independent, traditional antisemitism accounts for more than twice as much content as antisemitism tied to Israel or Zionism.
OHPI does not concede the opposing point either. The report says that many who claim “anti-Zionism isn’t antisemitism” are seeking a free pass to engage in antisemitism, of whatever form, without consequences. It also defines Israel-related antisemitism using the IHRA standard: applying antisemitic tropes to Israelis or Zionists, denying the Jewish people the right to self-determination, comparing Israel to the Nazis, and holding Jews collectively responsible for Israel’s actions.
Most of the traditional material falls into a catch-all category of slurs and medieval accusations, which often makes up more than half of the content collected on a given platform. OHPI notes that the juicebox emoji has become a pejorative for Jews. Users post it under photos of Jewish people, possibly to avoid detection by automated moderation. On Gab and Telegram, traditional antisemitism consistently makes up more than 80% of the antisemitic material collected and sometimes exceeds 90%.
LinkedIn’s Israel problem
LinkedIn is the exception. On the professional networking site, antisemitism related to Israel and Zionism consistently accounts for more than 70% of the antisemitic material collected, sometimes more than 80%. LinkedIn and Reddit together supplied 15% of OHPI’s data but 30% of all Israel-related antisemitism. On some LinkedIn samples, content denying the Jewish people’s right to self-determination made up more than a third of all antisemitic material on the platform.
Oboler told J-Wire that LinkedIn informed Australia’s Royal Commission on Antisemitism and Social Cohesion that it does not recognize this form of antisemitism. The report urges LinkedIn to recognize that “Zionist” is used as a code word for Jew.
Threads is no refuge
Many users moved from X to Meta’s Threads after Elon Musk bought Twitter, expecting a cleaner environment. OHPI found that Threads recorded more antisemitic items per eight hours of monitoring than X in February and March 2026, and the same number in April. On Threads, the mix is shifting toward Israel-related hate: traditional antisemitic items fell from 124 per eight hours in February to 83 in September, while Israel-related items rose from 71 to 119.
Nazis in the comments
Incitement to violence, including support for Nazism, is rising on mainstream platforms, mostly in comment sections. Violence-related content made up 25.5% and 26.5% of the antisemitic material collected on TikTok in April and May 2026, and 27.6% on the neo-Nazi-heavy Gab in June and July. On YouTube, violence-related antisemitism now exceeds Holocaust-related content. Most Holocaust denial is now found on mainstream platforms rather than fringe sites.
Meta behaved while the commission was watching
Antisemitism on Facebook, Instagram, and Threads dropped during June and July 2026, when the Royal Commission held hearings on social media and Meta executives testified on July 6. The drop did not last, and levels rebounded in the following months. TikTok, YouTube and LinkedIn executives also testified, but their platforms showed no similar decline. Oboler acknowledged the timing could be coincidental. OHPI is calling on Meta to restore the automated removal systems it switched off in January 2025.
The data also captured the online response to the December 14, 2025 Bondi Beach Hanukkah massacre, in which 15 people were murdered by gunmen inspired by ISIS. Antisemitism on Facebook, Instagram, TikTok, Gab, and Reddit rose between December and January, a period that also included the government’s January 8 announcement of the royal commission.
Erasing October 7
The report opens with the online promotion of pro-Palestinian events held on October 7, 2026, the third anniversary of the massacre. OHPI documented promotions in Sydney, Melbourne, Wollongong, across the United States, and in London and Hannover that portrayed the terrorists as martyrs or fallen soldiers. A Sydney event used the poppy, the symbol of Australian war sacrifice, and described October 7 as the day “Palestine fought back.” A London student march was promoted with the red Hamas triangle. Most comments on these posts opposed the events. Those that supported them often glorified the attack or claimed it was an Israeli “false flag.”
A global pattern
OHPI’s findings match other data on post-October 7 antisemitism. The Combat Antisemitism Movement’s Antisemitism Research Center documented 13,339 antisemitic incidents worldwide between October 7, 2023 and October 1, 2025, a period CAM CEO Sacha Roytman Dratwa called the largest surge in antisemitism in modern history. In the first 18 weeks of 2026, CAM counted 2,543 incidents worldwide, an average of 141 a week. CAM’s monitoring also showed how fast online conspiracy theories now spread after an attack on Jews. After the attempted hijacking of Flydubai Flight 1073 in September, conspiracy theories claiming Jewish or Israeli involvement appeared within 67 minutes of the emergency signal, and more than 4,000 posts used the term “false flag.”
In the United States, the ADL recorded 9,354 antisemitic incidents in 2024, the highest since it began tracking in 1979. For the first time, a majority were tied to Israel or Zionism. Its 2025 audit counted 6,274 incidents, a one-third drop driven largely by quieter campuses, though still higher than any year before October 7. The ADL’s 2024 survey of online harassment found that 41% of Jewish American adults had changed their online behavior to avoid being recognized as Jewish.
Recommendations
OHPI makes 10 recommendations. They include government funding for continued monitoring, regulation with penalties large enough that companies cannot simply absorb them, fines, eventual blocking for platforms that refuse to cooperate with inquiries, and more human trust-and-safety staff working alongside AI. The report concludes that platforms cannot be trusted to solve the problem out of goodwill.
“When a platform doesn’t recognise something, that’s what allows it to flourish,” Oboler said.
