Following the horrific events in Washington, DC, last week, here are some of the steps we’ve taken to protect the conversation on our service from attempts to incite violence, organize attacks, and share deliberately misleading information about the election outcome. It’s important to be transparent about all of this work as the US Presidential Inauguration on January 20, 2021, approaches.
Updated our coordinated harmful activity policy
We’ve been clear that we will take strong enforcement action on behavior that has the potential to lead to offline harm. Given the violent events in Washington, DC, and increased risk of harm, we began permanently suspending thousands of accounts that were primarily dedicated to sharing QAnon content on Friday afternoon.
Many of the individuals impacted by this updated enforcement action held multiple accounts, driving up the total number of accounts impacted. Since Friday, more than 70,000 accounts have been suspended as a result of our efforts, with many instances of a single individual operating numerous accounts. These accounts were engaged in sharing harmful QAnon-associated content at scale and were primarily dedicated to the propagation of this conspiracy theory across the service.
Our updated enforcement on QAnon content on Twitter, along with routine spam challenges, has resulted in changes in follower count for some people’s Twitter accounts. In some cases, these actions may have resulted in follower count changes in the thousands.
As stipulated in this policy that we announced ahead of the 2020 US election, accounts that have Tweeted or Retweeted associated content will continue to be subject to limited visibility across search, replies, and on timelines and are prohibited from being recommended to others by Twitter. It’s important that these types of accounts — that are not predominantly engaged in sharing this material — can see different perspectives in the open public conversation that Twitter uniquely provides.
Our teams are discussing ways we can empower research into QAnon and coordinated harmful activity on Twitter.
Escalated enforcement measures for our Civic Integrity Policy
During the past several weeks, misleading and false information surrounding the 2020 US presidential election has been the basis for incitement to violence around the country. We took action on these claims in line with our Civic Integrity Policy.
Now that the results of the election have been officially certified by Congress, we updated our Civic Integrity policy on Friday to aggressively increase our enforcement action on these claims. The updated policy provides details about how we enforce against violations of this policy, including repeated sharing of Tweets that receive warning labels. Ultimately, repeated violations of this policy can result in permanent suspension.
Deployed tech to surface potentially harmful Tweets for urgent human review
Our teams are continuing to aggressively deploy technology to surface potentially harmful Tweets for human review in an effort to take action as quickly as possible on violative content. Using this combination of technology and human review helps our teams work at scale during this critical time. We continue to update these tools as terminology and behaviors evolve on Twitter.
Limited engagement on labeled Tweets
On Tuesday, we limited engagement by no longer allowing any Tweets labeled for violations of our civic integrity policy to be replied to, Liked or Retweeted. People on Twitter are still able to Quote Tweet to share this content with additional context or their own perspective.
Blocked violative keywords from Search and Trends
We want Trends to promote healthy conversations on Twitter. This means, at times, we may prevent certain content from trending. There are rules for Trends, and if we identify Trends that violate these rules, we’ll take enforcement action.
Since last week, we've prohibited certain terms from surfacing in Trends and Search based on the following Twitter Rules:
We will also continue to prioritize reviewing and adding context to Trends. Our goal is to help people see what’s happening while ensuring that potentially confusing trends are presented with context.
Fought spam and challenged potentially inauthentic accounts
It is against the Twitter Rules to engage in spamming behavior, including bulk, aggressive, or deceptive activity. That’s why we routinely deploy anti-spam challenges to accounts to fight this behavior and protect the public conversation. During these challenges, account owners must verify their authenticity through a variety of measures, such as reCAPTCHA or providing a functional email address.
As always, while accounts are undergoing these challenges, they’re temporarily removed from follower counts. This, along with our updated enforcement around coordinated harmful activity, means some people may notice drops or fluctuations in their follower count.
Ahead of the inauguration, we’ll continue to monitor the situation, keep open lines of communication with law enforcement, and keep the public informed of additional enforcement actions.