The popular online service decided to fight the publication of inappropriate, graphic content. Check out the report YouTube published!

YouTube, for the first time in its history, published a report on compliance with community guidelines and the fight against harmful content. Over three months (from October to December 2017), more than 8.2 million videos that violated the terms of service and contained extremist, violent, or graphic content were deleted from the platform. It is worth adding that as many as 6.6 million of these clips were identified by artificial intelligence.

 

8,000,000 videos removed by YouTube! — illustration 1

 

Nearly 76 percent of the content was removed before it was viewed by any internet user. Meanwhile, among the videos that users managed to see before they were deleted, more than half recorded up to 9 views. The share of content viewed from 10 to 99 times is decreasing, as is the share of those that reached more than 100 views before being removed from the service.

 

8,000,000 videos removed by YouTube! — illustration 2

 

8,000,000 videos removed by YouTube! — illustration 3

 

In the period reported by YouTube, users flagged more than 9.32 million — that is exactly how many videos (in the period reported by YouTube) users judged to be content that violated the terms of service. Most reports of this type came from India, the United States, and Brazil.

 

8,000,000 videos removed by YouTube! — illustration 4

 

8,000,000 videos removed by YouTube! — illustration 5

 

What were the flagged video materials about? For the most part, their subject matter was related to sexual content (30.1%), spam and misleading information also reached a high reporting rate (26.4%), and additionally the pool included images of brutality and hate (15.6%).

 

8,000,000 videos removed by YouTube! — illustration 6

 

“Machine learning systems allow us to flag content for review at a larger scale, helping us remove millions of harmful videos before they are ever viewed - YT emphasizes in the statement announcing the publication of the report. - Our investment in machine learning, intended to speed up the removal of content that violates our policies, is beginning to bring the expected results both in higher-risk areas where a small number of videos is published, such as violent extremism, and in those with a large number of videos, such as spam.”

“Last year we committed that by the end of this year we would increase to 10,000 the number of people working on identifying content that violates our guidelines across Google services. At YouTube, we have also filled most positions related to the additional roles necessary to achieve this goal. We have also hired - full-time - experts specializing in issues related to violent extremism and counterterrorism as well as human rights, and we have expanded regional teams of experts.” - YouTube reports.