Meta Rolls Out Natural-Language Controls for Facebook Feed Recommendations
A new 'your algorithm' menu lets users define their interests and type custom feed instructions using large language models.
Meta is rolling out a new personalization tool for Facebook that allows users to adjust recommendation algorithms using natural-language prompts, Engadget reported .
The update introduces a "your algorithm" menu accessible from individual feed posts and Reels. Within the interface, users can add topics to their list of interests, mark subjects they want to avoid, or write custom text prompts describing the content they prefer to see.
According to Meta, the feature relies on large language models integrated directly into its recommendation ranking architecture. Rather than relying solely on engagement metrics like clicks and watch time to infer user preference, Meta stated that LLMs allow ranking systems to process explicit context and intent expressed through everyday phrasing.
The change mirrors a similar customization tool deployed across Instagram earlier this year. However, the Instagram version carried explicit operational bounds: it did not permit users to bypass algorithmic ranking to request only posts from accounts they follow, nor did it offer controls to reduce advertisement frequency.
The roll-out comes as social platforms continue grappling with user fatigue around low-quality automated recommendations and engagement bait in algorithmic feeds.
Sources
Written by
The Company Wire
Inside the companies building what’s next. Reporting on startups, technology, funding and the people shaping them.
