Facebook’s new AI-powered recommendation controls give users a more direct voice in their feeds. But choosing what interests you is not the same as controlling how a platform distributes information.
Facebook users have spent years teaching the platform what they like without necessarily intending to. Every video watched, post shared and comment written helps determine what appears next. Now Meta is introducing a more direct approach. On October 13, the company announced Your Algorithm, an AI-powered feature that allows people to describe, in ordinary language, what they want to see more or less frequently in their Facebook Feed and Reels. As TechCrunch reported Tuesday, the change moves Facebook toward a recommendation system users can actively instruct rather than influence primarily through their behavior. The promise is greater personalization. The unresolved issue is how much control those instructions actually provide.
From Watching Your Behavior to Asking What You Want
Meta says users can access Your Algorithm through the three-dot menu on Facebook posts, where they can manage topics and express their interests. The company is also expanding its Interested and Not interested controls. Users may receive follow-up questions asking why they liked or disliked something, with their answers used to adjust recommendations during the same session. According to Meta’s October 13 announcement, large language models help interpret the context and intent behind those responses. Someone interested in entrepreneurship, for example, could ask to see more small-business financing coverage without having to spend days interacting with similar posts first.
That represents a change in how people communicate with recommendation systems, but not a complete replacement of how those systems work. Meta says Facebook already evaluates thousands of behavioral signals when deciding what to show users. Written preferences will become another input alongside viewing history, engagement and other information the platform collects. Facebook previously introduced Show More and Show Less controls in 2022, using machine learning to interpret those signals. The new feature makes the feedback more conversational and potentially more specific. It does not mean users can independently determine every post that appears in their feeds.
The Algorithm Remains a Business Decision
The development also reflects the competing interests behind social media personalization. Users want relevant content and fewer unwanted recommendations. Creators and publishers want their work to reach interested audiences. Meta operates a business that depends heavily on advertising and continued engagement. A system that responds more accurately to users’ preferences could serve all three interests, but those objectives will not always align. Giving someone more influence over the subjects they encounter does not necessarily explain why one publisher receives greater distribution than another or how commercial considerations affect the ranking of content.
For publishers, the implications are particularly significant. Facebook has long been a distribution channel where audience access depends partly on decisions made inside an algorithm that outsiders cannot fully inspect. More explicit preferences could help specialized publications reach people who genuinely want their coverage. A reader asking for more reporting about higher education, labor markets or technology policy might encounter relevant journalism more frequently. But publishers will still depend on Meta’s interpretation of those interests, its ranking priorities and the visibility it assigns to different types of content. User-directed personalization does not automatically translate into publisher-directed distribution.
More Choice Does Not Necessarily Mean More Transparency
Meta has already been extending similar controls across its platforms. Instagram introduced Your Algorithm for Reels in December 2025 and expanded personalization to Feed and Explore in June 2026. Facebook’s adoption suggests the company sees conversational recommendations as a broader direction for social media. The interface is becoming easier to understand because people can express preferences in their own words. Yet the underlying decisions remain more complicated: which posts qualify as relevant, which sources are prioritized and how competing recommendations are ranked are still determined by the platform.
There is also a question about information diversity. A person who repeatedly requests more coverage of a particular subject or perspective may receive a feed that better reflects their interests while encountering fewer unfamiliar ideas. That possibility is not unique to Facebook’s new feature; recommendation systems have long raised concerns about narrow information environments. But more precise personalization creates another reason to examine whether platforms help users discover information beyond what they already know they want.
Facebook’s latest update gives people a more direct way to influence an experience that has often felt outside their control. That is a meaningful product change, particularly for users frustrated by irrelevant recommendations. Whether it becomes a meaningful shift in power will depend on what happens after those preferences are submitted. Users can now tell Facebook what they want to see. Facebook still decides what deserves to be shown.