Facebook Stories AI system

UPDATED MAR 12, 2025
The content you see on your Facebook Stories 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 Stories

When you view and interact with Facebook, one of the underlying AI systems delivers stories to you. This content appears at the top of your Feed and comes from people and Pages on Facebook.
Note: Facebook Stories may also include advertisements. That content is not powered by the AI we describe in this system card.
How Facebook Stories works
The AI system behind Facebook Stories automatically shows you stories from people or Pages by predicting what you’re most likely to be interested in. Here’s how it works:
  1. Gather stories
    First, the system gathers all relevant stories shared by people or Pages within the last 24 hours. It then attempts to remove stories that may go against our Community Standards.
  2. Make and analyze predictions
    Next, the system takes all stories that are available to you and applies predictions about stories you’ll find most relevant and valuable. The system keeps these stories and removes the rest.
  3. Rank stories
    The system then ranks this smaller set of stories based on how likely you are to interact with each story.
  4. Apply additional rules
    Finally, the system applies rules to ensure a balanced mix of stories from people and Pages.
How to customize what you see
Your experience on Facebook Stories 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.
Report
If you see a story that you think goes against Facebook’s Community Standards, you can report it.
Unfollow
You can unfollow a friend, Page or person to stop seeing stories from them.
See oldest stories first, before they disappear
To see stories from those you follow in chronological order, click "Stories" in the Feeds tab. Learn more.
Mute a story
To stop seeing a story, you can mute it. After doing so, you won’t see any more stories from the person or Page that produced the story unless you unmute them.
Share
You can share a story by sending it to someone via Messenger, adding it to your own story or adding it to your Page’s story.
How the AI delivers content to you
We want you to see content you enjoy and find interesting. To achieve this, the AI system has models that help it make predictions about content you'll find most relevant and valuable. These prediction models use underlying input signals to help select content you're most likely to engage with.
Below are some of the significant predictions–and input signals that inform them–that we use in this AI system.
The predicted number of stories you will watch in an author's collection
Signals influencing this prediction include:
  • The total amount of time users have spent watching the author's stories
  • The total number of stories that people have viewed from the author compared to the total number of stories created by the author
  • The total number of stories that people have viewed from the author
  • How many stories in the author's collection you haven't seen compared to the number of stories in the author's collection
  • How many stories in the author's collection you haven't seen
How likely you are to click an author's story that you've seen at the top of your Feed
Signals influencing this prediction include:
  • The total amount of time people have spent viewing the author's stories
  • The total number of stories that people have closed by navigating right in the author's collection
  • How many times you've viewed stories
  • How many times the story has been clicked to be viewed in full screen, compared to the number of times the story has been seen at the top of the Feed
  • How many times people have clicked to view the story in full screen
How likely you are to finish an author's story collection and move to the next author's story collection
Signals influencing this prediction include:
  • How many times you've skipped a story by clicking "X" to exit a collection of stories
  • How many stories you've skipped by clicking left compared to how many stories you've skipped by clicking "X" to exit a collection of stories
  • How many stories and collections of stories from different authors you've viewed
  • How many stories you've skipped by clicking left compared to how many stories you've skipped by clicking right
Probability of user consuming more than 3 cards in a story bucket
Signals influencing this prediction include:
  • number of times a user has consumed a story card in the last 28 days
  • How many stories in the author's collection you haven't seen
  • How many unique Stories has the user consumed in the last 28 days that are not their own
  • umber of times a user has consumed a story card that is oesp in the last 28 days
  • unique media consumed by the user in the last 28 days
How likely you are to close an author's collection of stories that you've started to watch
Signals influencing this prediction include:
  • How many times you've closed a story by clicking "X" to exit a collection of stories
  • How many stories in the author's collection you haven't seen
  • How many stories and collections of stories from different authors you've watched
  • The total amount of time people have spent viewing the author's stories
How likely you are to swipe right on the current story to move to the next author's story collection
Signals influencing this prediction include:
  • How many stories you've skipped by clicking left compared to how many stories you've skipped by clicking right
  • How many times you've skipped a story by clicking "X" to exit a collection of stories
  • How many stories you've skipped by clicking left compared to how many stories you've skipped by clicking "X" to exit a collection of stories
  • How many stories and collections of stories from different authors you've viewed
The predicted amount of time you will spend watching each new story added to an author's collection
Signals influencing this prediction include:
  • The total number of stories that people have closed by navigating right in the author's collection
  • The average number of stories from the author that people have watched
  • How many of the author’s Stories that users have watched have been played and advanced automatically
  • How many photos and videos are in the author's collection of stories
  • The average amount of time that people have spent viewing the author's stories
How likely you are to react to a story in an author's collection that you've started to watch, in ways such as clicking to like
Signals influencing this prediction include:
  • The total number of times people have clicked the arrow to send replies to the story's author
  • How many times the user has clicked the arrow to send a reply to the author of the Story compared to the number of Stories from the author the user has watched
  • How many times you've clicked the arrow to send a reply to the story's author
  • How many times the user has clicked the arrow to send a reply to the author of the Story compared to the number of Stories from the author the user has clicked to open
  • How many times you've clicked the arrow to send a reply to the story's author compared to the number of stories from the author you've watched
How likely you are to watch stories in an author's collection for longer than the average amount of time that other people, with the same number of unseen stories as you have, spend watching stories
Signals influencing this prediction include:
  • How many photos and videos are in the author's collection of stories
  • The average number of stories from the author that people have watched
  • The total number of stories from the author that have advanced automatically to the next story as people have watched them
  • How likely the user is to reply with text via Messenger to a Story in an author's collection that the user has started to watch
  • the average amount of time users spent viewing stories from the author
How likely you are to reply with text to a story in an author's collection that you've started to watch
Signals influencing this prediction include:
  • The total amount of time people have spent viewing the author's stories
  • Where data privacy laws permit, how many times you've exchanged messages on a mobile device with the story's author
  • How many text replies to stories you've sent
  • Where data privacy laws permit, how many times you've communicated with the story's author such as through chat messages
  • The total number of stories that people have closed by navigating right in the author's collection