Recommendations may amplify content, so it’s important they’re surfaced responsibly. Promoting healthy conversations is one of Twitter’s core principles; “freedom of speech is a fundamental human right — but freedom to have that speech amplified by Twitter is not.” While our enforcement philosophy empowers people to understand different sides of an issue by allowing many forms of speech to exist on our platform, we also work hard to prevent the amplification of harmful content on Twitter.
We have several ways of preventing potentially harmful or offensive content and accounts from being amplified, including using machine learning technology, reviewing user reports, and other tools. So, while harmful or offensive content and accounts may still exist on the Twitter platform, they may not be eligible for amplification in the following ways:
As a recommendation in Home timeline from someone you don’t follow
As part of the top section of Replies in Tweet details or conversations
As a recommendation on the “For you tab” in Explore
As a push notification
Content that can’t be recommended will still be available on Twitter to people who follow the Tweet author and on the Tweet author’s profile.
During important events including, but not limited to, certain crises and elections, the likelihood, immediacy, severity, and potential for harm may increase significantly. We take additional steps in these cases to prevent potentially misleading or harmful content from being inadvertently amplified. For example, during events such as civic processes and certain crises, we may proactively recommend informative messages or updates to add context to disputed or potentially misleading narratives that emerge. This includes creating Twitter Moments and launching prebunks, which proactively feature informative messages or updates designed to counter emergent misleading narratives.
Content ineligible for recommendations
Because there is a range of enforcement options for certain potential violations, Twitter does not always take down potentially problematic content. However, we strive to keep the following categories of content out of our recommendations:
Content from state-affiliated media accounts or the accounts of governments that are known to restrict access to the open internet during armed interstate conflicts.
Content that includes information known to be obtained through a hack and does not come from websites (specific articles or whole domains) that comment on or distribute materials in the course of some additional purpose, such as journalist coverage or commentary.
Content that violates any of the Twitter Rules, but has been left on the platform due to the public-interest exception.
Content that promotes the use of regulated substances or weapons.
Content that is deemed marginally abusive and is ineligible for amplification under our safety policies, including our Abusive Behavior and Hateful Conduct policies.
Harmful misleading information, including violations of the Civic integrity policy, COVID-19 misleading information policy, Synthetic and manipulated media policy, and any other misleading information policies listed at help.twitter.com/misinfo.
Content that automated systems have determined may violate the Twitter Rules, but that has not yet been reviewed by a human and/or may have been identified in error.
Accounts ineligible for recommendations
Because there is a range of enforcement options for certain potential violations, Twitter does not suspend all potentially problematic accounts. However, we strive to keep the following categories of accounts out of our recommendations:
Accounts that recently violated the Twitter Rules.
Accounts sharing harmful misleading information, including violations of the Civic integrity policy, COVID-19 misleading information policy, and Synthetic and manipulated media policy, and any other misleading information policies listed at help.twitter.com/misinfo.
Accounts that are thought to be engaged in spammy behavior.
Accounts that engage in coordinated harmful activity.
Accounts that contain graphic violence or hateful imagery in Twitter profile elements.
Accounts labeled as state-affiliated media or which belong to states that limit access to free information and are engaged in armed interstate conflict.
Where can I find recommendations?
We share recommendations on various areas on Twitter including Notifications, Topic Landing Pages, Explore, Spaces Tab, and the Home timeline. Below, we share more about recommendations on the Home timeline, Explore, and Notifications.
Home shows Tweets from accounts and Topics you follow, as well as other recommended content. This means you will sometimes see Tweets in your Home timeline from accounts you don't follow. We recommend Tweets to you based on a variety of signals including, but not limited to, accounts you follow, Topics you follow, Tweets you like, Tweets liked by those in your network, and accounts followed by those in your network. You may also see content such as promoted Tweets or Retweets from accounts you follow in your Home timeline.
We want you to understand why you’re seeing certain recommendations in your Home timeline. We do this in various ways, including by adding context to the Tweet you see, such as the name of the Topic, or who in your network follows or has Retweeted the Tweet author.
Learn more about the Home timeline and how you can control what you see in Home.
Explore catches you up on a range of subjects, tailored to your interests. This is where Twitter shows you what’s happening, right now. The more you use Twitter, the more personalized the Explore tab gets. Explore tabs may vary based on your location. Typically, you’ll see these sections under Explore:
For you: Trends, Events, Topics, and Moments that Twitter thinks you'll enjoy most, based on your activity.
Trending: what’s happening in the world. Trends displayed on the Trending tab represent what's trending within a specific geographic area, and are not personalized for every account.
News, Sports, COVID-19, Entertainment tabs: other top stories or news related to different subjects. Availability of these tabs may vary based on your location.
More about trends
Trends are meant to help you discover emerging topics of discussion on Twitter, and are determined by algorithms that identify conversations that are popular now, rather than topics that have been popular over a period of time. One input used by the algorithm to rank and determine Trends is the number of Tweets relating to a given topic. Algorithmically, Trends and hashtags are grouped together if they are related to the same topic. For instance, #MondayMotivation and #MotivationMonday may both be represented by #MondayMotivation.
When you’re signed into your account, Trends are tailored for you based on your account location. However, you have the ability to see what’s trending in other locations by changing your location setting.
Learn more about Explore and Trends.
Notifications help you stay in-the-know about the things you care about by sending you alerts. You can choose between notifications related to you and your Tweets and recommended notifications. From the Notifications timeline, you’ll be able to see which of your Tweets have been liked, the latest Retweets (of your Tweets), Tweets directed to you (replies and mentions), your new List followers, recommended notifications for you based on your activity on Twitter, as well as your new account followers. In addition to these notifications, we also elevate content that we think you’ll be most interested in and contributes to the conversation in a meaningful way, such as content that is relevant, credible, and safe as described above.
Notifications are personalized to each Twitter account and you can choose which notifications you want to receive for either your web browser or your mobile device. In addition to choosing notifications preferences, you have other options within the settings menu to filter the notifications you receive. These options include: quality filter, muted words, preferences, and advanced filters.
Quality filter, when turned on, it filters lower-quality content from your notifications, for example, duplicate Tweets or content that appears to be automated — it does not filter notifications from people you follow or accounts you’ve recently interacted with. You have the option to turn this on or off in your notifications settings.
Mute allows you to mute notifications for particular words and phrases you’d like to avoid seeing in your notifications. You can learn more about muting words and phrases here. You can also mute notifications for accounts you'd like to avoid seeing notifications from. This includes muting accounts you follow or don't know. For muted accounts you follow, replies and mentions by the muted account will still appear in your notifications timeline. You can learn more about muting accounts here.
Advanced filters allow you to disable notifications from certain types of accounts you’d like to avoid. In addition, if your account receives a lot of sudden attention, we may insert a notification in your Notifications tab inviting you to adjust these filters to give you more control over what you see.
Preferences allow you to choose which types of notifications you’d want to receive via email and on your mobile device. Some of these notification types include Mentions and replies, Topics, News, Moments, and Spaces.
Learn more about notifications.