Skip to main content

YouTube says it will recommend fewer videos about conspiracy theories

YouTube says it will recommend fewer videos about conspiracy theories


Taking steps to reduce the spread of misinformation

Share this story

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

Illustration by Alex Castro / The Verge

YouTube said on Friday that it would promote fewer videos containing misinformation and conspiracy theories, a potentially significant step to reducing the service’s potential to drive viewers toward extremist content. In a test limited to the United States, the service said it would stop recommending what it calls “borderline content” — videos that come close to violating its community guidelines but stop just short.

While YouTube said the change would affect less than 1 percent of videos that are available on YouTube, the sheer volume of content available on the service suggests the effect could be significant.

“We’ll begin reducing recommendations of borderline content and content that could misinform users in harmful ways — like videos promoting a phony miracle cure for a serious illness, claiming the earth is flat, or making blatantly false claims about historic events like 9/11,” the company said in a statement.

The move comes amid sustained criticism that YouTube has unwittingly pushed extremist videos for years. In a widely read op-ed last year, researcher Zeynep Tufekci called YouTube “the great radicalizer.”

“It seems as if you are never ‘hard core’ enough for YouTube’s recommendation algorithm,” Tufekci wrote. “It promotes, recommends and disseminates videos in a manner that appears to constantly up the stakes. Given its billion or so users, YouTube may be one of the most powerful radicalizing instruments of the 21st century.”

More outrageous videos tend to perform well

By default, YouTube enables an autoplay feature that serves users new videos as soon as the one they are watching finishes. These rabbit holes often lead viewers to videos that have received lots of “engagement,” including views, comments, and shares. More outrageous videos tend to perform well by these measures, and so they have been served up by the YouTube algorithm for years.

On Thursday, BuzzFeed published a new investigation of YouTube’s recommendation engine and found that a new account clicking on a nonpartisan political video would be watching extremist content within just six videos.

“Compounding this issue is the high percentage of users who say they’ve accepted suggestions from the Up Next algorithm — 81%, according to Pew,” BuzzFeed wrote. “One in five YouTube users between ages 18 and 29 say they watch recommended videos regularly, which makes the platform’s demonstrated tendency to jump from reliable news to outright conspiracy all the more worrisome.”

YouTube’s move leaves many questions remaining about which kinds of videos will be considered borderline. In an interview, the company suggested that users should read its publicly posted community guidelines for guidance on which kinds of videos may come close to receiving a dreaded strike.

Moderators will train the system to recognize borderline videos

Enforcement of the new policy will fall to a mixture of human moderators and machine learning systems, the company said. Moderators will train the system to recognize borderline videos, and then the system will be deployed to review videos automatically and decide whether they are eligible for promotion.

The opaque nature of the process could help YouTube prevent users from gaming the system. But it is also highly likely to lead to frustration with users who feel that their videos have been unfairly labeled as borderline content.

Videos that are labeled borderline will not be removed from the site, the company said. If a user subscribes to a channel that features borderline content, those videos will still be eligible to be shown in the subscriber’s recommendations. “We think this change strikes a balance between maintaining a platform for free speech and living up to our responsibility to users,” the company said in its blog post.

If the test is successful, YouTube plans to roll out the changes more broadly.

YouTube is not alone in paying fresh attention to questions about content that becomes popular by walking the fine line between banned and permitted. In November, Facebook said that it would act to halt the spread of “sensationalist and provocative” content. “I believe these efforts on the underlying incentives in our systems are some of the most important work we’re doing across the company,” CEO Mark Zuckerberg wrote in a blog post.