Facebook's "People You May Know" feature uses a sophisticated AI-driven algorithm to suggest potential friends based on multiple factors related to your social connections and personal information. The core elements Facebook considers include:
- Mutual friends: The most common and significant factor is shared connections with other users.
- Shared affiliations: This includes attending the same school, working at the same company, or belonging to the same Facebook groups.
- Contact information: Facebook uses phone numbers and email addresses uploaded by you or your friends to find matches.
- User interactions and network exploration: The algorithm looks at your existing network and common interests to identify possible connections.
- Other undisclosed factors: Facebook keeps the full workings of the algorithm confidential, but it may incorporate additional data points such as wireless signals or indirect connections.
The system continuously updates and ranks suggestions using AI to deliver a personalized list of people you might know
. Facebook also collects data not only from your own contacts but potentially from your friends’ contacts, which can create suggestions even if you haven't uploaded your own information. This has raised privacy concerns because the feature can reveal connections based on data shared by others, not just yourself
. Importantly, simply visiting someone's profile does not directly cause them to appear in your suggestions. The algorithm relies more on network and contact data rather than profile views
. Users can manage their experience by dismissing suggestions, controlling who can send them friend requests, or adjusting privacy settings to limit visibility in these recommendations
. In summary, Facebook's "People You May Know" feature leverages a complex AI system analyzing mutual friends, shared affiliations, uploaded contacts, and other data to predict and suggest people you might know, while keeping the exact algorithm details proprietary
