Meta collects and infers information about you from what other people do
Meta collects information about you based on other people's activity, and it also infers things about you from others' activity. Its examples: it might suggest a friend to you through Facebook's People You May Know or Instagram's Discover People if you both appear on a list of contacts or connections that someone uploads, and it takes into account whether your friends belong to a group when it suggests you join it.
Supporting quotations
We collect information about you based on others' activity.
We also infer things about you based on others' activity. For example:
We might suggest a friend to you through Facebook's People You May Know or Instagram’s Discover People if you both appear on a list of contacts or connections that someone uploads.
We take into account whether your friends belong to a group when we suggest you join it.
Qualifications & exceptions
The friend-suggestion example depends on another person uploading a contact or connection list that includes you both.
Not stated here
Footnote 12's further examples are expanded elsewhere in this Policy; only the examples in this section are used here.
Inference & Profiling · Company practice
Ads are chosen using your profile, on- and off-Product activity, content, AI chats, inferences and connections
To decide what ads to show you and others, Meta uses information it has about you, including: your profile information; your activity on and off its Products, including information received through cookies and similar technologies, according to your settings; content you create or interact with across Facebook and Instagram; your interactions with its AI at Meta features; things it infers about you, like topics it thinks you may be interested in; and information about your friends, followers or other connections, including their activity or interests. It also uses de-identified, aggregated or anonymized information to show you ads, and points to a separate help page about ways it shows ads, including using machine learning. Information with special protections that you choose to provide is not used to show you ads.
Supporting quotations
To decide what to show you and others, we use information we have about you, including:
Your profile information
Your activity on and off our Products, including information we receive through cookies and similar technologies, according to your settings
Content you create or interact with across Facebook and Instagram
Interactions with our features that are part of
Things we infer about you, like topics we think you may be interested in
Information about your friends, followers or other connections, including their activity or interests
about some of the ways we show you ads that we think may be interesting to you, including using machine learning.
We use information with special protections you choose to provide for these purposes, but not to show you ads.
Qualifications & exceptions
Off-Product activity received through cookies and similar technologies is used 'according to your settings'.
The list is introduced with 'including', so it is not exhaustive.
Information with special protections you choose to provide is not used to show you ads (stated in the provide-and-improve part of the section).
Not stated here
The linked help page on machine learning in ads is not supplied.
Meta uses your interests, activity, connections, device and partner data to tailor Meta AI responses and actions
Meta uses information it has, including about your interests and preferences, to make responses and actions from its AI at Meta features more relevant. The categories named are your activity and information you provide; friends, followers and other connections; app, browser and device information; and information from partners, vendors and third parties. Meta's example: when a user asks Meta AI for a dinner recommendation, Meta AI uses her location-related information, plus that she recently told Meta AI she loves Italian food and has been liking Instagram posts about modern fine dining, to suggest a nearby upscale Italian bistro, and then offers to book a table. The example shows information from Instagram activity and earlier Meta AI chats being combined to personalize an AI response.
Supporting quotations
We use information we have, including about your interests and preferences, to help make responses and actions from our features that are part of
more relevant, such as:
Your activity and information you provide
Friends, followers and other connections
App browser and device information
Information from partners, vendors and third parties
For example, Anna is looking for a place to eat and asks Meta AI to recommend a restaurant for dinner. Meta AI uses Anna's
To make the restaurant suggestion even more relevant, Meta AI also uses additional information including that Anna recently told Meta AI she loves Italian food, and that she has lately been liking Instagram posts featuring modern fine dining. By using her interests, location and profile, Meta AI is able to respond to Anna with a relevant suggestion of a highly-rated, upscale Italian bistro nearby. Meta AI then also offers to book Anna a table at that restaurant.
Qualifications & exceptions
The categories are introduced with 'such as'; the list is not exhaustive.
The restaurant scenario is Meta's illustration, not a description of every AI interaction.
Inference & Profiling · Company practice
Facebook check-ins and Page likes can drive Instagram ads and inferred interests; you can manage ad topics
Meta's ad examples: a user checks into a local bakery on Facebook, so Meta might later show him Instagram ads for other local bakeries; another user likes several Facebook Pages about famous musicians, so Meta thinks he is interested in 'music' and might show him an ad for a local record shop or online music publication. He can manage his interests in his ad topics setting.
Supporting quotations
Your activity on our Products
For example, William checks into a local bakery on Facebook, so later we might show him ads on Instagram for other local bakeries.
Topics we think you might be interested in
For example, Jon likes several Facebook Pages about famous musicians, so we think he has an interest in "music."
Based on this interest, we might show him an ad for a local record shop or an online music publication.
Jon can manage his interests in his
Qualifications & exceptions
The examples use 'might'; they illustrate possible ad selection, not a guaranteed outcome.
Inference & Profiling · Company practice
Your friends' activity, such as liking an Instagram account, can shape the ads Meta shows you
Meta's example for using your connections in ads: a user's friend likes an Instagram account for a local art fair, and based on her friend's activity Meta might show her an ad for the art fair.
Supporting quotations
Your connections
For example, Fiona’s friend likes an Instagram account for a local art fair. Based on her friend’s activity, we might show Fiona an ad for the art fair.
Qualifications & exceptions
The example uses 'might'; it illustrates a possible outcome.
Inference & Profiling · User right or choice
Advertisers target audiences by interests Meta infers, such as Page likes; you can view and remove your interests
Meta's example of general demographics and interests: a bike shop wants its ad to reach people in Atlanta interested in cycling, and Meta decides whether someone fits that audience based on things like whether they liked a Page about bikes; people in that audience could then see the ad. You can see the 'interests' assigned to you in your Ad preferences and remove them if you want. The policy also points you to Ad preferences to manage the ads you see on Meta Products.
Supporting quotations
General demographics and interests
For example, a bike shop creates a Page on Facebook and wants to place an ad to reach an audience of people in Atlanta interested in cycling. We determine whether someone fits in this audience based on, for example, whether they liked a Page about bikes. Then people in that audience could see the bike shop's ad.
You can see the “interests” assigned to you in your
and remove them if you want.
Manage the ads you see on our Products
Visit your
Qualifications & exceptions
Page likes are one example of how audience fit is decided ('for example').
Inference & Profiling · Company practice
Meta's systems automatically analyze data about you and others to infer your interests for feeds, ads and suggestions
In the part of the policy setting out why and how it processes information, Meta says its systems automatically process information collected and stored about you and others to assess and understand your interests and preferences and give you personalized experiences on and across Meta Company Products, in accordance with its terms. This is how it personalizes features and content (such as Facebook Feed, Instagram Feed, AI at Meta features and Stories); personalizes the ads people see; and makes suggestions for you, such as people you may know, groups or events, or topics to follow, on and off its Products. Its ads system prioritizes which ad to show you based on the audience advertisers want to reach, then matches the ad to people who might be interested.
Supporting quotations
Why and how we process your information
The categories of information we use, and why and how information is processed, are set out below:
Personalizing Meta Products
Our systems automatically process information we've collected and stored associated with you and others to assess and understand your interests and your preferences and provide you personalized experiences on and across the Meta Company Products in accordance with our terms. This is how we:
Personalize features and content (such as your
and Stories);
Personalize the ads people see, and
Make suggestions for you (such as people you may know, groups or events that you may be interested in or topics that you may want to follow) on and off our Products.
Our ads system prioritizes what ad to show you based on what audience advertisers want to reach. Then we match the ad to people who might be interested.
about how we choose the ads that you see.
Qualifications & exceptions
Processing uses information associated with you and with others.
Suggestions can be made off Meta Products as well as on them.
Not stated here
The footnote [86] on how the ads system works is outside this passage.
Inference & Profiling · Company practice
Meta predicts your ad interests, matches advertisers' audiences and ranks ads with machine learning
Meta's ads system works this way: it uses your information to determine the ad interests most relevant to you, trying to predict them from the information it has, including choices you have made in Ad preferences. Advertisers say which audience they want to reach, which can include age, gender, location or interests (for example, women in Dublin who like golf). Its system automatically reviews all potential ads where you might fit the advertiser's audience, and decides which to show based on things like whether an ad best achieves the advertiser's goal and is most relevant to you, using algorithms and machine learning to order ads. Factors include ad quality, the advertiser's desired audience and budget, and how likely Meta thinks you are to react to the ad based on things like your activity on Meta Products. It uses this process to choose the ad you see on Meta Company Products.
Supporting quotations
How our ads system works
Our ads system works in the following way:
We use your information to determine the ad interests most relevant to you. We try to predict your ad interests based on the information we have, including any choices you've made in Ad preferences.
Advertisers tell us which audience they want to reach based on who they think will be most interested. This can include things like age, gender, location where advertisers want to show their ad or interests, such as women, based in Dublin, who like golf.
We review the potential ads that we could show you to identify the most relevant ones. Our ads system automatically reviews all the potential ads where you might fit the audience that advertisers want to reach. For example, you may see potential ads because you match the age range and interest category. We determine which ones to show you based on things like whether it best achieves the advertiser’s goal and is most relevant to you. To do this, we use algorithms and machine learning to order the ads we may show you. We use different factors such as the ad quality, the advertiser's desired audience and budget. We also consider how likely we think it is that you will react to the ad based on things like your activity on Meta Products.
All of this information helps us show you the most relevant ads. When you see an ad on Meta Company Products, we use this process to choose the most relevant ad to show you.
Qualifications & exceptions
Your Ad preferences choices are one input to the interest predictions.
Advertiser budget and goals are ranking factors alongside relevance to you.
Not stated here
The linked help page on ad ranking is not supplied.