Instagram Builds AI Feature to Let Users Tune Reels Algorithm

Instagram is developing an AI-powered feature that gives users more direct control over how the Reels algorithm serves them content, according to a report from AfroTech. The move signals a continued shift at Meta toward user-tunable recommendation feeds — and it carries real implications for creators who have built distribution strategies on algorithmic reach.

Details on the rollout remain thin. The report does not specify a launch date, a testing market, or whether the feature is in internal testing, limited beta, or wide release. What it does confirm is direction: Instagram wants viewers, not just the ranking model, to shape what appears in their Reels feed.

That direction is not new for Meta. Instagram already offers users the ability to mark interests, snooze suggested posts, and indicate "Not Interested" on individual Reels. CEO Mark Zuckerberg has repeatedly framed AI-driven personalization — including user feedback loops — as the core of the platform's content strategy. This feature appears to push that logic further, reportedly using AI to translate user preferences into algorithmic adjustments rather than relying on discrete taps and settings menus.

For creators and social media managers, the practical question is what this does to reach. When users can tune their feeds more aggressively, content that misaligns with a viewer's stated preferences faces a harder path into the recommendation pool. The likely result: interest alignment becomes more valuable than raw engagement bait. Accounts that publish across loosely related niches may see recommendation reach fragment further, while tightly focused accounts could benefit from users who actively opt into their topic area.

There is also an ads angle, though the report does not address it. If preference signals become richer and more explicit, Meta's ad targeting could eventually draw on the same data. That remains speculation — no ad-product changes have been announced alongside this feature.

How much control users actually want is an open question. Instagram's own experiments with chronological feeds and favorited accounts showed that many users say they want more control but rarely reconfigure their settings after the first week. An AI-mediated version could lower that friction by inferring preferences from behavior rather than requiring manual curation. Whether that inference works well enough to change retention or session time is exactly what Meta will be watching.

Practitioners should treat this as a signal, not a crisis. No algorithm change has shipped. No reach data exists yet. The reasonable preparation is the same as for any personalization push: sharpen niche definition, watch audience retention metrics closely, and avoid assuming that engagement hacks will survive a system where the viewer holds a bigger dial.

Expect concrete rollout details — test markets, user-facing controls, and any creator-facing metrics — to surface through Meta's official channels in the coming months.