Open the Reels tab, tap a tiny icon that looks like two lines with hearts, and you’ll get something Instagram has seldom offered before: a readable explanation for why certain short videos keep appearing on your feed. The feature, called “Your Algorithm,” launched in the U.S. this week and lets people see an AI‑generated summary of their recent interests and then tell Instagram to show more or less of specific topics.
It’s a small change in the interface but a potentially big one in how platforms and people negotiate attention. "We use AI to summarize your interests based on your activity," Tessa Lyons, Instagram’s vice president of product, told Good Morning America. "And if we get it wrong, you can remove those interests from your algorithm."
Why this matters
For years, users have complained that algorithmic feeds feel opaque — you know the app knows something about you, but you can't see the evidence. Instagram’s new dashboard tries to pull back the curtain: it lists the subjects the system thinks you're into, gives you chips to add or subtract topics (think "Horror movies," "Chess," or "College football"), and even offers a one‑tap way to share that snapshot to your Story.
That combination — visibility plus direct control — is what sets this apart from previous nudges. Lance Ulanoff, editor at large for TechRadar, told Good Morning America that the move helps demystify the core part of a social app: the algorithm. For many users, understanding what the algorithm is actually doing can change how they interact with the app.
How it works (and what it doesn’t)
The feature is AI‑driven: Instagram summarizes your recent activity and populates the interface automatically, though you can edit the topics it proposes. Initially it lives in Reels and is rolling out in English globally soon; Instagram says it plans to expand similar controls to Explore and other parts of the app.
This isn’t an escape hatch from personalization. The tool refines the existing recommender rather than turning personalization off. If you want a full reset, older tests have suggested Instagram might add a reset button — but "Your Algorithm" is designed for fine tuning, not demolition.
Privacy and second‑order effects
There are two threads worth watching. One is privacy: yes, you can see what Instagram thinks you like, but the summary itself is generated by the company’s models and can be shared publicly by you. That raises questions about how summaries are stored and whether revealing them changes social dynamics — will people curate their algorithm publicly as another form of self‑presentation?
The other is the larger landscape of AI that personalizes digital life. As companies push features that analyze and act on user data, we’ve seen related debates about search and data access elsewhere: Google’s growing experiments with Gemini‑powered features that dive into Gmail and Drive highlight similar tradeoffs between convenience and systemic access to personal content. You can read more about that development in the piece on Gemini’s Deep Research.
And this isn’t happening in a privacy vacuum. Meta’s hardware and software projects have generated their own privacy conversations — for instance, recent firmware changes to Ray‑Ban Meta Glasses raised questions about where device data goes and who can access it. Those conversations echo here: handing users a peek at algorithmic reasoning helps, but it doesn’t erase the underlying data flows that power these recommendations. See the update on Ray‑Ban Meta Glasses firmware and privacy for more on those concerns.
A practical note for users
If you want to try it, look for the hearts‑on‑sliders icon in Reels. Expect the initial list to reflect what you’ve watched and liked recently; if you see stuff that’s off—remove it. If you’re worried about lock‑step personalization, experiment: add topics you wouldn’t normally see to diversify the recommendations. Instagram says it will add more nuanced feedback tools for Reels and Facebook in the coming months, and it’s testing a similar control for Threads.
This is not a magic fix. It won’t stop algorithmic amplification of sensational content, nor will it change ad targeting directly. But for people tired of feeling at the mercy of inscrutable recommendation engines, the feature is a rare gesture in the direction of user agency — a screwdriver, not a sledgehammer, to pry something open that used to be sealed shut.