Instagram’s Algorithm Directed Users to Antisemitic Content Within Days, Study Finds
Platform Directed Users to Hate Content They Never Sought Then Refused to Remove It When Reported
Instagram’s recommendation algorithm steered ordinary users from mainstream wellness content to antisemitic conspiracy theories within three days, according to a study released June 3 by the Antisemitism Research Center, the research arm of the Combat Antisemitism Movement.
The Lioness Writes is a fully reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Annual subscription fee is on sale through 6/8 for $70…a bit over $1/day.
Researchers built two test accounts — one simulating a wellness-focused user, one a fitness-focused user — and engaged only with mainstream self-improvement material. Neither account searched for or interacted with extremist content of any kind.
The wellness account was served 59 classifiable videos over the study period. More than 32% were coded or explicit antisemitic content. The fitness account received 71 classifiable videos, of which 24% were identified as antisemitic.
Antisemitic content appeared during the first browsing session, before either account had built any meaningful engagement history.
Recommended videos included conspiracy theories linking the Israel-Iran war to a plot against the American food supply, rants about Jewish financial power, and claims that the Rothschild family was behind the sinking of the Titanic.
“You don’t have to search for antisemitic content to find it on Instagram,” said Oliver Marks, research director at the Antisemitism Research Center. “When platforms optimize for engagement without sufficient safeguards, they can end up amplifying hate to vast audiences.”
Meta’s Moderation Rollback
The findings come more than a year after Meta overhauled its content moderation policies across Instagram, Facebook, and Threads in January 2025.
The changes ended third-party fact-checking in the United States and replaced proactive enforcement with a user-reporting model. Meta CEO Mark Zuckerberg described the shift as “a trade-off.”
Following the rollback, antisemitic comments on the Facebook pages of Jewish members of Congress increased nearly fivefold, according to a May 2025 Anti-Defamation League report.
Reporting System Fails to Remove Content
Between January and February 2026, ADL researchers submitted 253 pieces of violative content through Instagram’s standard user-reporting system — the same tool Meta directs users to when they encounter harmful material.
The flagged content included posts tied to white supremacist networks, accounts linked to designated foreign terrorist organizations, and vendors selling merchandise bearing Nazi symbols.
Instagram removed 11 accounts and 8 posts. In 20 cases, the platform told the ADL it did not have the capacity to review the submissions.
The overall removal rate was 7%.
A Meta spokesperson said after the report’s release that some of the flagged content “did not violate our policies.” The ADL had given Meta two weeks to act before publishing its findings.
In February 2026, Meta’s oversight board confirmed the company’s AI moderation system was only being applied to content that users had already manually flagged — not content the platform was scanning for on its own.
Threads
Meta’s Threads platform faces similar concerns, though no comparable algorithmic study has been conducted on the app.
Accounts displaying Nazi imagery appeared on Threads within 24 hours of the platform’s 2023 launch, according to the Tech Transparency Project. Nick Fuentes, a white supremacist banned from Instagram, publicly announced he had created a fake account on Threads at launch and encouraged his followers to do the same.
In April 2026, Meta’s director of content policy pointed to the company’s prohibition on Holocaust denial as evidence of its commitment to combating online hate. Jewish advocacy groups noted it had taken four years of sustained pressure to get Meta to classify Holocaust denial as a policy violation.
Unsurprisingly, Meta did not respond to a request for comment on the Antisemitism Research Center’s findings.
SOURCES
ARC/CAM full study: https://combatantisemitism.org/studies-reports/algorithmic-escalation-from-self-improvement-content-to-antisemitism-on-instagram/
ADL report on Instagram moderation failures: https://www.adl.org/resources/press-release/alarming-proliferation-antisemitism-and-extremism-instagram-new-adl-report
Jewish Insider on ADL findings: https://jewishinsider.com/2026/04/adl-study-instagram-extremist-content-meta-moderation/
ADL on Meta policy rollback: https://www.adl.org/resources/article/metas-hate-policy-rollback-linked-increased-antisemitism
Wow, this got me so angry!
Disgusting. I don't have an account with them. Good thing I'm thinking