|
Getting your Trinity Audio player ready...
|
The Antisemitism Research Center (ARC) by CAM recently conducted a systematic observational study on Instagram to examine whether the social media platform’s recommendation algorithm unintentionally creates exposure pathways between mainstream self-improvement content and antisemitic material.
Two fresh persona accounts simulated ordinary users in distinct aspirational ecosystems — wellness and biohacking; and fitness and discipline. Both followed only mainstream creators.
Across three days of 45-minute sessions, both accounts were served content that escalated, without any active user intent, from clean eating advice and fitness motivation to conspiracy theories and explicit antisemitic content, including translated Nazi propaganda.
Read the full report, “Algorithmic Escalation: From Self-Improvement Content to Antisemitism on Instagram,” HERE.
Key Findings
The wellness persona was served 59 classifiable videos over three days, of which 32.2% were classified as either coded antisemitism or explicit antisemitism, including 9 explicit antisemitic videos in a single 45-minute session on day three.
The fitness persona recorded 71 classifiable videos over three days, of which 24% were classified as either coded or explicit antisemitism, including 17 coded or explicit antisemitic videos in a single 45-minute session on day three.
The findings are significant for three reasons —
Speed: Coded antisemitic content was served to the wellness account on day one, within its first session, before the account had built any meaningful interaction history.
Volume: On day three, 31% of the wellness account’s recorded content and 18% of the fitness account’s recorded content was classified as explicit antisemitism.
Convergence: Despite different starting points, both accounts were independently routed toward the same antisemitic narrative clusters, tropes, scapegoating logic, and, in several cases, the same specific pieces of content. This suggests the issue is not only a community-level phenomenon but a deeper failure in the algorithmic architecture.
All documented content is archived and linked HERE in full, allowing independent verification of all findings.
The ARC had previously uncovered networks of more than 80 AI-generated fake “rabbis” promoting antisemitic tropes on Instagram, along with 63 such accounts on YouTube and 49 on TikTok.
Meta, Instagram’s parent company, took action to stop the proliferation of the fake “rabbi” content exposed by CAM, and CAM now urges similar responsiveness to the issues raised in this latest report.
Additional information regarding the study’s methodology can be found HERE and HERE.
Read more:
Fake AI ‘Rabbis’ Spread Antisemitic Tropes in Coordinated TikTok Campaign, New CAM Study Reveals
Fake AI ‘Rabbis’ Push Antisemitic Conspiracies to Millions Online
Instagram’s Algorithm Feeds Antisemitic Content to Millions, New CAM Research Finds
