Now, the platform has announced several updates to its efforts against extremism — an issue that likely goes beyond morality for YouTube given that a massive ad boycott kicked off this spring after marketers discovered that spots were running against videos espousing extremist ideals and violence.
Perhaps most interestingly, YouTube said it will soon place videos that don’t necessarily violate its community guidelines — but that have been flagged by users for hate speech — “in a limited state.” Videos that contain controversial religious or supremacist content, for instance, will be able to viewed under a warning but won’t be recommended, monetized, or include comments, suggested videos, or likes. This new capability will roll out to YouTube on desktops in coming weeks, with mobile access to follow.
In the realm of machine learning, YouTube says it has implemented new technology to identify and remove terrorism-related content “in a scalable way,” and that it has already seen “some positive progress.” Over the past month, YouTube says its systems have doubled the number of videos removed for extremism, and that 75% of these videos were taken down before they received a single human flag.
Finally, YouTube says that it has begun working with 15 additional NGOs — including the Anti-Defamation League and the No Hate Speech Movement — to help it better identify content that’s used to radicalize and recruit terrorists. The company also plans to hire more people to review and enforce policies around extremist content, and is taking a proactive approach to fighting extremism via its Creators For Change program, which supports YouTube channels that discuss social issues and promote tolerance.
“We are confident we can continue to make progress against this ever-changing threat,” the company wrote in a blog post today. “We look forward to sharing more with you in the months ahead.”