Meta expands its AI-powered “Your Algorithm” feature from Instagram to Facebook and Facebook Reels and says it is using LLMs more in its ranking systems
Meta's bringing Instagram's “Your Algorithm” feature to Facebook, and using LLMs more to rank content.
Context & Ripple Effects
Meta has been building AI deeper into distribution rather than treating it as a standalone assistant: its 2023 disclosures detailed recommendation systems across Facebook and Instagram, while the 2025 Reels refresh prioritized fresher, more relevant recommendations. The Facebook rollout extends an Instagram control model across more of Meta’s feed surfaces.
The move also connects ranking with Meta’s broader AI interface strategy. Its June AI Mode launch used public Facebook posts, Groups, and Reels to answer searches, making the quality and steerability of content selection more consequential across both discovery and search-like experiences.
First-order effects
- Facebook Feed and Reels users gain a direct way to shape the topics their recommendations emphasize, while Meta applies LLMs more extensively to the ranking decisions behind those feeds.
- Creators and publishers on Facebook and Reels face distribution that is influenced both by Meta’s LLM-driven relevance judgments and by audiences’ expressed topic preferences.
Second-order effects
- Meta can carry a common personalization-control pattern from Instagram into Facebook and Reels, giving its product teams a shared user signal to refine recommendations across the apps.
- The 2025 emphasis on fresher Reels content is reinforced by a system that combines recency and relevance with user-selected interests, raising the value of timely, clearly categorized posts for distribution.
Third-order effects
- If Meta continues pairing user controls with LLM ranking, feed personalization shifts from a largely opaque recommendation system toward one where users can actively tune the inputs without taking over curation themselves.
- The broader platform competition moves toward AI-mediated distribution: the differentiator is not only what content is available, but how precisely a platform can match it to stated and inferred interests at scale.
The trend: Social platforms are turning recommendation feeds into adjustable AI systems, combining explicit user preferences with increasingly capable ranking models.