Video network YouTube responded to critics who have long called on the Google company to clean up its recommendation engine, and not offer conspiracy videos in suggested plays.
In a blog post posted Friday, YouTube said it will begin to reduce recommending “borderline content and content that could misinform users in harmful ways — such as 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 most high-profile conspiracy theorist, Alex Jones, was banned by YouTube in 2018, along with Apple, Stitcher, Facebook and others over concerns about his content, including videos that questioned whether the killings at Sandy Hook were real.
So, if YouTube viewers were searching for news accounts of an event like Sandy Hook, they could have seen InfoWars pop up after a string of videos from CBS, ABC and the like. The YouTube algorithm picks videos based on your viewing and search histories in an auto play mode, with suggestions for others videos to watch as well.
This isn’t the first time YouTube has tried to take the conspiracy video problem. In March 2018, it said it would put links to Wikipedia about events in text-boxes around “widely accepted events, like the (1969) moon landing.”
More: YouTube tries to solve its conspiracy problem with Wikipedia. Some critics cry foul.
More: 7 days from fringe to mainstream: How a conspiracy theory ricocheted around the web
YouTube was responding to public backlash after its trending videos tab gave a wide platform to clips that shared false or misleading information about breaking national news events, like the high school shooting in Parkland, Florida, or the 2017 mass shooting in Las Vegas.
YouTube says this algorithm shift will apply to less than 1 percent of the content on YouTube, but that “limiting the recommendation of these types of videos will mean a better experience for the YouTube community.”
The network emphasized that the change will affect recommendations of what videos to watch, not whether a video is available on YouTube.
“As always, people can still access all videos that comply with our Community Guidelines and, when relevant, these videos may appear in recommendations for channel subscribers and in search results.”
YouTube receives over 400 hours of videos uploaded to the site every minute, so many questionable videos fall through the cracks. Critics have long called for more human oversight, and YouTube has responded by saying it would hire as many as 10,000 people to hunt for offensive videos. Additionally, the network has touted its use of machine learning to get a better handle of keeping bad stuff off the network.
Videos remain up until they are “flagged” by members of the community for not adhering to YouTube’s guidelines, such as showing violence, sexual content or in the case of the recent Florida high school shootings, false information.
An example of the conspiracy issue. In February 2018, when student survivors of a high school shooting in Florida began speaking out, a video that suggested the students were “crisis actors,” hired by Democrats and gun control advocates, rose to the top of YouTube’s Trending section, attracting 200,000 views. Once the media caught onto it, YouTube pulled it down.
This change in policy relies on a combination of machine learning and real people, YouTube says. “We work with human evaluators and experts from all over the United States to help train the machine learning systems that generate recommendations. These evaluators are trained using public guidelines and provide critical input on the quality of a video.”
YouTube had resisted making wholesale changes to its algorithm due to concerns for free speech. The network said the change “strikes a balance between maintaining a platform for free speech and living up to our responsibility to users.”
The changes will take effect for now only in the United States and will eventually be rolled out to other countries.