How does X detect purchased followers with its new anti-spam AI?

X’s newest anti-spam AI is designed to identify follower purchases through pattern recognition rather than simple account checks. In 2025, the system evaluates behavioural signals across millions of profiles simultaneously. Purchased followers tend to behave differently from real users, and the AI relies on those behavioural discrepancies. When a batch of new followers arrives, the system monitors how they interact with your content. Real users usually scroll, like posts, reply occasionally, or at least view your profile. Purchased followers typically show no behaviour at all beyond hitting the “follow” button.

Another major detection method is traffic source analysis. If your account suddenly gains followers from unrelated regions, identical devices, or clusters of IP ranges linked to automated networks, the AI flags it as suspicious. It also looks at the timing of follows — automated followers often appear in tight intervals, which the system marks as non-human patterns.

Additionally, X’s updated AI links follower authenticity to account history. Profiles created recently, or ones that follow thousands of accounts with zero posts, raise immediate red flags. When enough of these signals accumulate, the system may reduce your account’s visibility, remove suspicious followers, or trigger additional verification checks.

Overall, X’s 2025 anti-spam AI is highly behaviour-driven, focusing on unnatural growth patterns rather than individual follower profiles.