Instagram says it will use machine learning to find and remove inauthentic likes, follows, and comments from accounts using third-party apps to boost popularity
Instagram is fighting back against automated apps people use to leave spammy comments or follow then unfollow others in hopes of growing their audience.
Context & Ripple Effects
This announcement extends a campaign Instagram has been running since it began cracking down on automation services like Instagress in mid-2017 (that earlier crackdown targeted the same follow/like/comment bots). What changes now is the method: rather than acting only against the bot services themselves, Instagram is applying machine learning directly to engagement signals on its own platform.
The move also fits a pattern visible weeks earlier, when Instagram deployed machine learning to scan photos for bullying and route them to human moderators — engagement fraud is simply the next surface for the same detection stack. Five years later the problem persists at scale, prompting the bulk fake-follower cleanup tools of late 2023 (flag-and-delete tooling).
First-order effects
- Users running third-party growth apps face direct removal of purchased likes, follows, and comments, gutting the visible metrics those services sell.
- Influencers and brands whose audience numbers were inflated by bot activity see follower counts and engagement rates drop once the cleanup runs.
Second-order effects
- Third-party growth-automation vendors lose their product's value proposition and are pushed toward either shutting down or shifting tactics that evade detection.
- Advertisers and marketers get cleaner engagement data, which raises the relative value of authentic reach and pressures rivals like Twitter and Facebook to match equivalent enforcement.
Third-order effects
- If the pattern holds, platform trust becomes an arms race between ML detection and evasion tooling, with platforms progressively taking over moderation functions — flagging, removal, even user-facing cleanup — that were previously left to users or outside services.
- Engagement authenticity turns into a compliance-grade metric: as platforms certify their own counts, inflated-audience intermediaries (bot farms, growth hacks) become structurally unviable rather than merely against the rules.
The trend: Social platforms are shifting moderation from reactive policy enforcement against bot vendors to proactive machine-learning policing of engagement signals themselves, a line running from the Instagress ban through this announcement to today's bulk fake-follower tools.