CAM Report: Instagram Algorithm Surfaces Antisemitic Content to Millions

A new report from the Combat Antisemitism Movement (CAM) alleges that Instagram's recommendation algorithm surfaces antisemitic content to millions of users, a finding the organization summarized in a headline statement: "Instagram's Algorithm Feeds Antisemitic Content to Millions."

The research, surfaced this week through CAM's own channels and indexed by Google News, points to the platform's feed, Reels, and Explore recommendation mechanics as the distribution vehicle. CAM's claim shifts responsibility from individual hate-posting accounts to the ranking system itself, a distinction with direct implications for how advertisers, agencies, and creators evaluate adjacency risk on Meta-owned properties.

What is CAM's claim?

CAM's headline finding is that the algorithm, not just bad-actor accounts, funnels antisemitic material into users' feeds. The full report text, methodology, sample size, and dataset were not included in the headline summary distributed by the organization. CAM has not, in the materials available, published a detailed methodology or named specific accounts, posts, or impression counts that would allow independent verification.

That gap matters for how the report should be read. CAM is an advocacy organization focused on tracking antisemitism, and its research output is best understood as investigative advocacy rather than peer-reviewed academic study. Meta has historically disputed similar amplification claims from outside groups, arguing that its systems reduce exposure to violating content rather than increase it.

What does this mean for marketers?

For brand advertisers, an algorithm-level finding changes the risk calculus. Most brand-safety stacks target accounts, hashtags, and post-level content. An amplification finding implies hate material can surface next to neutral creator content through Reels suggestions and Explore grids, where adjacency controls are thinner.

For creators, the implication runs in the same direction: a Reel can be recommended alongside content the creator did not produce and does not endorse, including hate material. This dynamic is not unique to antisemitism; amplification concerns have surfaced around misinformation, self-harm, and political content across Meta's properties.

What should practitioners watch next?

Three signals will indicate how seriously Meta treats the report:

• A public response from Meta, which typically arrives within days of an advocacy amplification report and usually combines disagreement with a policy-review statement.
• Whether CAM publishes the full methodology and dataset, which would let independent researchers test the claim.
• Whether the report surfaces in any pending regulatory filing, including EU Digital Services Act compliance reviews or U.S. congressional inquiries into platform safety.

CAM's finding joins a wider body of research alleging that engagement-optimized recommendation systems amplify harmful content at scale. The next data point for marketers and creators will be Meta's response and whether the report moves from advocacy publication into regulatory or legal proceedings.