CAM Study Flags Suspicious Amplification Around Account Hostile to Israel
The Combat Antisemitism Movement has identified strong signs of automation or false identities in nearly one quarter of the accounts it examined around @Wallface, an X profile that publishes content hostile to Israel. The finding puts the machinery of online amplification under scrutiny, alongside the messages being circulated.
Published on October 21, 2026, the study by CAM’s Antisemitism Research Center flagged 2,597 of 10,709 unique accounts, or 24.3%. Those are classifications produced by the center’s detection model. They describe a selected network of followers and people engaging with posts, rather than the entire audience for discussion of Israel.
What the researchers actually counted
The review took place on October 19 and 20. It covered 7,276 recent followers and accounts that reposted, replied to or quoted 139 posts. CAM said the information came from public data available through X’s official interface. Accounts appearing in several groups were counted only once in the overall total.
The largest flagged share was among recent followers, at 27.2%, followed by reposters at 23.7%. Commenters registered 13.9%, while accounts quoting posts registered 8.0%. That variation matters: following a profile, spreading its posts and commenting on them are different activities, and the results do not justify treating every interaction as equally suspicious.
Behavior is the test
CAM says its model assesses 32 signals, including posting bursts, account age, repeated text and follower patterns. Political views are excluded from the score. The study used a threshold of at least 75 on a scale of 100, looking for combinations of signals rather than condemning an account for one unusual habit.
This separates the subject of the investigation from the method used to examine it. A position hostile to Israel is not itself evidence of automation. Equally, a large visible audience does not establish that every profile represents a separate person. The distinction is essential when engagement figures become part of an argument about public opinion.
The investigation extends beyond one profile
CAM also found that 428 flagged accounts had appeared in its earlier investigations, with 416 flagged there as well. The overlap supports examining activity across multiple audiences. It does not identify a common operator or establish that the owners of prominent profiles commissioned the engagement.
An earlier CAM study, released on October 13, examined 10,015 accounts amplifying 230 widely shared posts by Mario Nawfal. It flagged 4,054, or 40.5%. The two studies used different samples, so their percentages should not be read as a direct ranking of users or a measurement of all activity on X.
X already has rules against artificial reach
X’s authenticity policy prohibits deceptive identities and coordinated activity intended to inflate the prominence of accounts or content. It also permits genuine collective expression and distinct accounts serving legitimate purposes. Its stated enforcement options include restricting reach, temporarily limiting features and suspending accounts.
For readers following Israel and Jewish affairs, the immediate implication is to distinguish popularity from corroboration. Thousands of reposts can distribute a claim without supplying any additional evidence for it. The new study offers a specific set of accounts for scrutiny, while the platform’s existing rules provide a framework for assessing deceptive conduct.
