Facebook Feed Recommendations AI system

UPDATED JUN 4, 2025
The content you see on your Facebook Feed Recommendations is selected, ranked and delivered to you by an artificial intelligence (AI) system. Within one AI system, multiple machine learning models work together to deliver your experience. These models and their input signals are dynamic and they change frequently as the system learns and improves over time.

Overview of Facebook Feed Recommendations

When you view and interact with Facebook, one of the underlying AI systems delivers suggested content to your Feed on the Facebook Home tab. This is content you may be interested in, including content from Pages or groups you may not follow. Sometimes you’ll also see suggestions for other content like reels, events and live video.
Note: Your Feed might also include connected content and advertisements. That content is not powered by the AI we describe in this system card. Connected content is described in Facebook Feed.
How Facebook Feed Recommendations works
The AI system behind Facebook Feed Recommendations automatically determines which content shows up in your Feed, and in what order, by predicting what you’re most likely to be interested in or engage with. These predictions are based on a variety of factors, including what and whom you’ve followed, liked or engaged with recently. Here’s how it works:
  1. Gather inventory
    The system gathers a portion of the public content available on Facebook, which may include photos, videos and links. This does not include content that goes against our Community Standards or that goes against our Recommendation Guidelines.
  2. Leverage signals
    Next, the AI system considers a variety of input signals about the content. These signals might include how you’ve engaged with similar content or your interests on Facebook.
  3. Make predictions
    From there, the system has models that help it make predictions about public content you’ll find most relevant and valuable.
  4. Rank content
    Finally, the system ranks content from the previous step. Content that the system predicts will provide more value for you is shown higher in your Feed. This step helps the system make content recommendations that more closely match your preferences.
How to customize what you see
Your experience on Facebook Feed Recommendations is personalized based on your activity, and you have options to control or customize what you see. Below, we describe how to do this with different in-product features. Options shown here may not be available to everyone.
Discover non-personalized content
To discover reels, videos and other content that is not personalized specifically to you, use the search feature in the Feeds tab. By doing so, your search results will be based on the search term you enter. Learn more.
Hide
You can hide a post. This action also helps to limit similar content from showing up in your Feed.
Interested/Not interested
Clicking “Interested” or “Not interested” on a post will temporarily increase or decrease the ranking score for a post and other posts like it.
Report content
If you see something you think goes against our Community Standards, you can report it to us. You can also report posts that you think are spam or false news.
How the AI delivers content to you
How long you are predicted to spend looking at a post
Signals influencing this prediction include:
  • The content type of the post, for example photo or video
  • The length of the video
  • The content type of the post, for example photo or video
  • How long have you been seen the triggered post
  • Time stamp of the session
Whether or not you will share a post
Signals influencing this prediction include:
  • How frequently user shares posts
  • How many times you've shared posts created by a specific person or page
  • The current time
  • How many times user shared posts related to a specific topic over the last 30 days
  • List of posts the user has shared in the past 180 days.
How likely you are to click “X” on a post
Signals influencing this prediction include:
  • How many recommended posts you have viewed in the last 7 days
  • How many recommended posts you have viewed in the last day
  • The total number of views the user has had in the past 14 days
  • How many recommended posts you have viewed in the last 14 days
  • How many posts you have viewed over the past 7 days
How likely you are to spend more time viewing a post
Signals influencing this prediction include:
  • How many video posts you have viewed, by topic
  • How long you have viewed the post while on a mobile device
  • How many seconds you viewed each post, by topic
  • How long you have viewed the post while on a mobile device
  • How long you have viewed the post while on a mobile device capturing longer duration.
How likely you will join a group
Signals influencing this prediction include:
  • The number of open groups that the user is member of
  • The number of closed groups that the user is member of
  • The number of pages that the user has liked in the last week
  • The number of pages that the user has liked
  • Topics that user is interested in based on group engagement
Whether or not you will click to like a post
Signals influencing this prediction include:
  • Number of likes the user has given to posts related to a specific topic or interest in the last 30 day.
  • Features of the post
  • The current time
  • Number of likes from the viewer to the actor of the post in the past 3 months
  • How often people like a post compared to how many times it's been shown, for each post creator.
How likely you will privately share a post on other platforms
Signals influencing this prediction include:
  • How many times a user has shared posts on Messenger in the past 1 month.
  • How many times a user has shared posts in the past 3 months
  • Owners whose post the user has shared on whatsapp in the last 180 days
  • How many times a user has shared posts in the past 1 week
  • How many times a user has shared posts in the past 1 month
How likely you are to take a screenshot of a post
Signals influencing this prediction include:
  • List of the owners the user has screenshot their post in the last 180 days
  • List of posts where the user has screenshot in the past 180 days.
  • The number of times the post has been saved
  • The number of times the post has been viewed.
  • How often user engaged with different types of ads based on content type.
How likely you are to click the rsvp button of a post
Signals influencing this prediction include:
  • The number of times you have interacted with the event post in the past 28 days
  • Your approximate distance from the event in a post
  • The number of times the post is clicked for recommended content on mobile devices
  • An approximate distance between your location and event venue
  • How often the post is clicked for recommended content on mobile devices
How likely you are to click 'Interested' if the interested/not interested options are presented to you below the post
Signals influencing this prediction include:
  • List of owners the user has clicked the interested control in last 180 days
  • List of posts the user has used "Interested" control in the past 180 days.
  • The current time
  • How many times user watched a video for 30 seconds or more in the last 30 days
  • The length of the video