YouTube Began Detailing Its ‘Four Rs’ Approach to Responsibility for Its Content

Remove is step one, and machine learning is a big part

YouTube's four Rs YouTube
Headshot of David Cohen

YouTube CEO Susan Wojcicki penned a blog post last week in which she reiterated the Google-owned video site’s goal of preserving openness through responsibility.

She wrote, “A commitment to openness is not easy. It sometimes means leaving up content that is outside the mainstream, controversial or even offensive. But I believe that hearing a broad range of perspectives ultimately makes us a stronger and more informed society, even if we disagree with some of those views. A large part of how we protect this openness is not just guidelines that allow for diversity of speech, but the steps that we’re taking to ensure a responsible community. I’ve said a number of times this year that this is my No, 1 priority. A responsible approach toward managing what’s on our platform protects our users and creators like you. It also means that we can continue to foster all the good that comes from an open platform.”

Wojcicki broke down YouTube’s approach toward responsibility into “four Rs”—remove, raise up, reduce and rewarding.

  • Remove: YouTube removes content that violates its policies as quickly as possible and strives to make those policies clearer and more effective.
  • Raise up: The video site raises up authoritative voices when people are looking for breaking news and information, especially during breaking news moments.
  • Reduce: Wojcicki said YouTube reduces the spread of content “that brushes right up against our policy line,” citing the site’s changes to its recommendations in January.
  • Rewarding: She reaffirmed YouTube’s commitment to rewarding “trusted, eligible creators.”

The video site went into detail on R No. 1, Remove, in a blog post Tuesday, saying that over the past 18 months, it has slashed views on videos that are eventually removed for policy violations by 80%.


YouTube said it has a dedicated policy development team to help it maintain a balance between preserving free expression and protecting its community, and that team consistently reviews, revises and clarifies its policies, adding that dozens of those updates have been implemented since 2018.

The video site provided an example: “Our hate speech update represented one such fundamental shift in our policies. We spent months carefully developing the policy and working with our teams to create the necessary trainings and tools required to enforce it. The policy was launched in early June, and as our teams review and remove more content in line with the new policy, our machine detection will improve in tandem. Although it can take months for us to ramp up enforcement of a new policy, the profound impact of our hate speech policy update is already evident in the data released in this quarter’s Community Guidelines Enforcement Report.”

YouTube added that the spikes it reported in removal numbers were caused in part by the removal of older comments, videos and channels that were previously permitted under its policies.

The company also detailed its use of hashes, or digital fingerprints, to detect copies of known content that violates its policies before they are ever viewed by people, and its contributions to shared industry databases on topics such as child exploitation and terrorism.

YouTube said more than 87% of the 9 million videos it removed during the second quarter of 2019 were initially flagged by its automated machine learning systems, with over 80% pulled before receiving a single view, adding that an update to its spam detection systems led to a rise of over 50% in the number of channels terminated for violating those policies.

The video site concluded, “We’re determined to continue reducing exposure to videos that violate our policies. That’s why, across Google, we’ve tasked over 10,000 people with detecting, reviewing and removing content that violates our guidelines. For example, the nearly 30,000 videos we removed for hate speech over the past month generated just 3% of the views that knitting videos did over the same time period.” David Cohen is editor of Adweek's Social Pro Daily.