OAKLAND, Calif. — Twitter took steps on Friday to slow the way information flows on its network, even changing some of its most basic features, as alarm grows that lies and calls for violence will sweep through social media in the weeks surrounding the presidential election.
The changes will temporarily alter the look and feel of Twitter. The company will essentially give users a timeout, for example, before they can hit the button to retweet a post from another account. And if users try to share content that Twitter has flagged as false, a notice will warn them that they are about to share inaccurate information.
Twitter also said it would add a label to claims about who won the election until it has been called by authoritative sources.
The steps announced on Friday are the most dramatic in a series of moves made by social media companies in recent months to stem the flow of misinformation in the lead-up to the Nov. 3 election, and are likely to draw the ire of Twitter’s most famous user, President Trump.
The companies are going to considerable lengths to avoid a repeat of the 2016 election, when Russian disinformation flowed unchecked on Facebook, Twitter and even YouTube, which is owned by Google. Facebook and Google have committed to banning political ads for an undetermined period after polls close on Nov. 3. Facebook also said a banner at the top of its news feed would caution users that no winner had been declared until news outlets called the presidential race.
In this election, much of the false news on the platforms has come from domestic sources and, in some cases, elected officials. That has forced the companies to walk a careful line between stopping false narratives because they go viral and have a real-world impact while countering arguments that they have become self-appointed censors.
“Twitter has a critical role to play in protecting the integrity of the election conversation, and we encourage candidates, campaigns, news outlets and voters to use Twitter respectfully and to recognize our collective responsibility to the electorate to guarantee a safe, fair and legitimate democratic process this November,” the Twitter executives Vijaya Gadde and Kayvon Beykpour said in a statement.
In recent weeks, Twitter has already added warning labels to lies posted by elected officials — flagging several of Mr. Trump’s tweets — and has cracked down on photos and videos that had been manipulated to deceive viewers. The company has not accepted political advertising for nearly a year.
Most of the latest changes will happen on Oct. 20 and will be temporary, Twitter said. Labels warning users against sharing false information will begin to appear next week. The company plans to wait until the result of the presidential election is clear before turning the features back on.
The Twitter executives said the “extra friction” on retweets will prompt users to add their own thoughts before they hit the button. If users decide they don’t have anything to add, they will be able to retweet after the prompt.
The change is likely to have a direct impact on Mr. Trump’s online activity. Since returning to the White House on Monday after a hospital stay to treat the coronavirus, he has been on a Twitter tear. On Tuesday evening, for example, he tweeted or retweeted posts from other accounts about 40 times.
The Trump campaign reacted angrily on Friday afternoon to Twitter’s changes, labeling them “extremely dangerous for our democracy.”
“After months of Big Tech censorship against President Trump, the unelected liberal coastal elites of Silicon Valley are once again attempting to influence this election in favor of their preferred ticket by silencing the President and his supporters,” Samantha Zager, deputy national press secretary for the Trump campaign, said in a statement.
The Biden campaign declined to comment.
Twitter’s changes ahead of the election are far more aggressive than those of its peers, and it could be sacrificing some of the ways it drives traffic to its service. Twitter will also disable the system that suggests posts on the basis of someone’s interests and the activity of accounts they follow. In their timelines, users will see only content from accounts they follow and ads.
“I believe the platforms are trying to throw everything at the wall to see what sticks here. It’s not clear what the right answer is but they are trying almost anything,” said Nate Persily, the co-director of the Stanford Cyber Policy Center.
Vanita Gupta, the president and chief executive of the Leadership Conference on Civil and Human Rights, said Twitter should consider making some of the changes permanent, given that elections are constantly occurring around the globe and Twitter has a duty to protect those elections as well.
“These are important steps but we’re going to be vigilant about how these play out in real time,” she said.
Twitter stopped short of shutting down its Trending Topics feature, a change that many critics say would do the most to fight misinformation because people can game the feature to promote false or misleading information. Instead, Twitter will expand its effort to fact-check and provide context to items that trend in the United States.
Over the last year, Twitter has slowly been stripping away parts of its service that have been used to spread false and misleading information, including misleading tweets from Mr. Trump.
That has led to a backlash from the Trump administration. Mr. Trump, who has 87 million followers on Twitter, has called for a repeal of legal protections Twitter and other social media companies rely on.
But Twitter’s fact-checking has continued. It recently began adding context to its trending topics, giving viewers more information about why a topic has become a subject of widespread conversation on Twitter. This month, Twitter plans to add context to all trending topics presented on the For You page for users in the United States.
“This will help people more quickly gain an informed understanding of the high-volume public conversation in the U.S. and also help reduce the potential for misleading information to spread,” Ms. Gadde and Mr. Beykpour said.
Twitter’s trends illustrate which topics are most popular on the service by highlighting content that is widely discussed. The trends often serve as an on-ramp for new users who are discovering how to find information on Twitter, but internet trolls and bots have often exploited the system to spread false, hateful or misleading information.
As recently as July, trending topics have been hijacked by white nationalists who pushed the anti-Semitic hashtag #JewishPrivilege and by QAnon, a conspiracy group that made the furniture company Wayfair trend on Twitter with false claims that the company engaged in child trafficking. The embarrassing episodes led critics to call on Twitter to shut down trends altogether.
Mr. Persily said the feature should have already been turned off. “What good does it really do?” he said.