Nearly one in four accounts that follow or amplify the X account @Wallface show bot-like signals, according to a study published on October 21 by the Antisemitism Research Center, part of the Combat Antisemitism Movement. The finding is a reminder for everyday users that follower counts and repost totals on X do not always reflect real people paying attention.
The researchers analyzed accounts on October 19 and 20, working from public, read-only X API data. Their sample covered 10,709 unique accounts: the 7,276 newest followers of @Wallface, about 14 percent of his roughly 51,000 followers, plus the accounts that amplified 139 of his posts, including 2,198 reposters, 1,202 commenters and 463 quoters. Some accounts appeared in more than one category.
Using its own detection model, the center flagged 2,597 of those accounts, or 24.3 percent, as likely bots or fake profiles. The flagged share was highest among the newest followers, at 27.2 percent, followed by reposters at 23.7 percent. Commenters were flagged at 13.9 percent and quoters at 8.0 percent, suggesting that accounts that simply follow or repost are more likely to show inauthentic signals than accounts that write a reply or add a quote.
The study also looked at timing. Among accounts amplifying posts from July, 15.9 percent were flagged. That share rose to 19.1 percent in August and 23.2 percent in September, before easing slightly to 22.2 percent in October. The researchers noted that the increase coincided with a sharp rise in @Wallface's posting activity. One individual post stood out in the opposite direction from its author's complaint: on October 19, @Wallface wrote that some of his posts had been mass-reported by bots in what he called platform manipulation aimed at silencing him, yet 21 of the 68 accounts that amplified that very post, or 30.9 percent, showed strong signals of inauthenticity.
How the score works matters for how readers should take the headline number. Each account was measured against 32 signals, including posting rate and bursts, odd-hour activity, duplicate text, account age, default avatars, digit-suffixed handles, follower patterns and generic or repetitive replies. No single signal counted on its own; only consistent combinations raised an account's score, which runs from 0 to 100. One flagged account, @Rajdhani108306, scored 99.9, after publishing approximately 171,000 posts since June 2026, an average of about 1,233 per day.
The center says its model, identified as X model v1.4.0, was built to undercount. Post content and ideology are not inputs, so an account is never scored as a bot for what it says, and any one measurable human behavior, such as original varied text, real conversations or a daily sleep gap, pulls a borderline account out of the bot-like group. Its signal weights draw on published research from DFRLab, Indiana University's Observatory on Social Media and the Cresci studies.
For regular X users, the practical lesson is modest but useful. Treat rapid follower growth and large repost counts with some skepticism, especially around accounts that post constantly, and look for varied, conversational activity before assuming an audience is entirely human. The researchers also found that 428 of the flagged accounts had appeared in their earlier investigations, which they say suggests inauthentic amplification is not limited to a single account or audience.