Twitter said it applied contextual labels to approximately 300,000 tweets for content that was disputed or potentially misleading during a two-week period covering the election. The figure comes as part of a wider post-mortem assessment of the company’s handling of political misinformation before and after Election Day.
Twitter has witnessed a wave of misinformation as users including President Donald Trump — who has nearly 89 million followers — and his allies have spread false and misleading claims about the election and its outcome. As of the morning of Nov. 7, 16 out of 43 of Trump’s post-election tweets had been labeled.
In addition to the labeling, Twitter (TWTR) said more than 450 of all the labeled tweets were also covered up by a warning message and were subject to sharing restrictions that limited how they could be retweeted. Roughly three out of four people who viewed those tweets did so after the labeling was applied, Twitter (TWTR) said in a blog post. The analysis focused on tweets about the US election from Oct. 27 to Nov. 11.
“We also want to be very clear that we do not see our job as done — our work here continues and our teams are learning and improving how we address these challenges,” wrote Vijaya Gadde and Kayvon Beykpour, who respectively lead Twitter’s legal and product teams. “We’ll be sharing a comprehensive report on the election early next year.”
The disclosure comes as Twitter rolls back certain preemptive policies that it put in place ahead of Election Day. Twitter said it found that removing recommendations to users for who they ought to follow had little meaningful impact on misinformation during the election, and the company will undo that change on Thursday.
The company said it will also relax some of the restrictions surrounding what trending topics users may see under a curated tab on its website labeled “For You.” During the election, Twitter said only topics that provided additional in-line context were permitted in that section of the site. That change is also being reversed.
One election-related change that Twitter will keep is an extra screen prompting users to quote tweet content instead of retweeting it. Twitter said its data showed that the limitation reduced sharing via quote tweets and retweets by 20%, and that it “slowed the spread of misleading information by virtue of an overall reduction in the amount of sharing on the service.” The company said it will continue to study the impact of the change and may revisit it in the future.
Twitter did not immediately respond to CNN’s request for data on how long it generally took Twitter to apply a label to misinformation, nor which accounts were primarily responsible for posting the tweets that ultimately got flagged.