The New York TimesApril 7, 2021 4:08:22 PM IST
It’s the endless battle for YouTube.
Every minute YouTube is bombarded with videos that go against its many guidelines, be it pornography or copyrighted material or violent extremism or dangerous misinformation. The company has perfected its artificial intelligence computing systems in recent years to prevent most so-called infringing videos from being uploaded to the site, but it remains under scrutiny for failing to curb the spread of dangerous content.
In an effort to demonstrate its effectiveness in finding and removing videos that violate the rules, YouTube on Tuesday revealed a new metric: the rate of views in violation. It’s the percentage of total YouTube views that come from videos that don’t meet your guidelines before the videos are removed.
In a blog post, YouTube said that infringing videos accounted for 0.16 to 0.18 percent of all views on the platform in Q4 2020. Or put another way, out of every 10,000 views on YouTube, 16-18 were for content. who broke YouTube’s rules and was eventually removed.
“We’ve made a lot of progress and it’s a very, very low number, but of course we want it to be lower,” said Jennifer O’Connor, director of YouTube’s trust and security team.
The company said its infringing view rate had improved from three years earlier: 0.63 percent to 0.72 percent in the fourth quarter of 2017.
YouTube said it was not disclosing the total number of times problematic videos had been viewed before they were removed. That reluctance highlights the challenges faced by platforms, such as YouTube and Facebook, which rely on user-generated content. Even if YouTube makes progress in capturing and removing banned content (computers detect 94 percent of problematic videos before they are viewed, the company said), total views are still a staggering number because the platform is so large.
YouTube chose to reveal a percentage rather than a total number because it helps contextualize how significant the problem content is to the platform as a whole, O’Connor said.
YouTube published the metric, which the company has tracked for years and expects to fluctuate over time, as part of a quarterly report that describes how it is applying its guidelines. In the report, YouTube offered totals for the number of objectionable videos (83 million) and comments (7 billion) it had removed since 2018.
While YouTube points to such reports as a form of liability, the underlying data is based on YouTube’s own decisions about which videos violate its guidelines. If YouTube finds that fewer videos are infringing and therefore removes fewer of them, the percentage of infringing video views may decrease. And none of the data is subject to an independent audit, although the company did not rule that out in the future.
“We start by just publishing these numbers and making a huge amount of data available,” O’Connor said. “But I wouldn’t take it off the table just yet.”
YouTube also said it was counting views generously. For example, a view counts even if the user stopped watching before reaching the objectionable part of the video, the company said.
Daisuke Wakabayashi [c.2021 The New York Times Company]