System Cards

The most significant change to Facebook’s resource center is the introduction of new system cards. This system consists of 14 cards:

• Feed

For around 500 posts, Facebook uses AI to establish a relevancy score and then ranks them in descending order. Because the system provides diverse information in the feed, a user should only see a few video postings in a row.

• Feed Ranked Comments

The AI ranks comments based on what it believes will be most relevant to each user. It does so by looking at other comments’ popularity and whether someone in their network published them.

• Feeding Suggestions

AI will predict which material users will most likely engage with based on characteristics such as groups they have recently joined and posts they have liked. It then utilizes this data to determine what content (for example, posts, reels, and live videos) to recommend.

• Reels

AI decides which reels to serve and in what sequence by predicting what a user will likely be interested in. It makes these predictions by looking at accounts the user has followed, liked, or engaged with recently.

• Stories

By anticipating what a person is likely interested in, the AI system automatically displays Stories from individuals or pages. The system also employs criteria to ensure users get diverse Stories content.

• Individuals you might know

AI attempted to identify persons of interest by examining criteria such as people who are friends with a user’s friends or people who belong to the same organizations as the user.

• Video

When users see and engage with Facebook Video, one of the underlying AI algorithms serves up a variety of video formats that may correspond to their interests. This content is present under the Video tab. It might involve films, music, games, or television series. The video content is that people may be interested in from creators they currently need to follow.

• Marketplace

One of the core AI algorithms recommends appropriate Marketplace listings when a user reads and interacts with Facebook, including the Facebook Marketplace feed. Users can, for example, view things for sale in categories such as household goods, pet supplies, and recreational goods. Other recommendations, such as merchants and information users may be interested in, may appear in their feeds.

• Notifications

AI selects and prioritizes alerts based on what it believes will be most relevant to the user. Meanwhile, previously seen signs showed up in the order they were received.

• Search

AI assigns a score to each possible search result based on how relevant that content is to a user, considering aspects such as content type. Based on this ranking, it will serve users the results in order of relevancy.

• Groups Feed

By assessing content based on relevancy, AI determines which posts show in the Groups feed and in what order.

• Individual Group Feed

AI analyzes what content users are most likely to engage with and then ranks it in their group feed based on relevancy. Relevancy considerations include what and with whom people have lately followed, liked, or interacted.

• Group Suggestion

Facebook’s AI will examine criteria such as groups that a user’s friends are members of and themes connected to items that a user may have recently engaged with, then use this information to identify other groups that may be of interest.

• Pages You Might Like

AI will recommend pages to follow based on recently liked by a user’s friends or pages that may be related to items and content the user has recently engaged with.

Meta’s Content Library and API

Meta’s new Content Library and API are another significant addition to Facebook’s resource center. The database is configured to include information from:

• Public Post
• Pages
• Groups
• Events
• A graphical user interface or a programmatic API may search, examine, and filter data.

However, under the present criteria, this tool has been designed particularly for researchers from certified academic and research institutions doing scientific or public-interest research. Researchers will need to apply to receive access to this data.

Personalizing the user experience

Facebook said that, in addition to increasing openness about its ranking algorithms, it also wants to provide people options to regain control over the information they see, such as the “Why Am I Seeing This?” feature. These technologies empower Facebook users to design their experiences and pick what they do and do not want to view. People may make adjustments by going to their Facebook Feed Preferences and Settings.

What did Facebook say?

On the Meta digital newsroom, Meta’s president of Global Affairs presented details about how AI Facebook ranks content and how users can decide what they see in the future. He stated:

• The AI systems increase the likelihood that the posts you see will be relevant and exciting to you. They’re also making it more transparent how you can improve your control over what you see on our applications, testing new settings, and making others more accessible. Furthermore, they are providing more thorough information to professionals so that they may better comprehend and analyze our systems.
• The AI systems anticipate how valuable content will be to you, allowing you to show it sooner. For example, sharing a post indicates that you found it fascinating; expecting whether or not you will share a post is one factor our systems consider.
• No prediction can accurately anticipate whether a post will benefit all. So they employ a range of predictions to get as near the proper material as possible, including some based on behaviour and others based on user input collected through surveys.
• By offering these products to academics early in development, they can collect helpful comments and produce the best solutions to satisfy their needs.