YouTube Community Guidelines Enforcement Caused Removal of 7.8 Million Videos
Without any doubt, YouTube is the biggest platform for sharing videos. As a part of Google Transparency Reports, YouTube publishes quarterly reports regarding the deletion of videos due to policy violations. This year in the July 2018 to September 2018 quarter, the company took down 7.8 million videos. Not only that but they took down nearly 1.7 million channels and over 224 million comments. And according to YouTube, machine learning continues to play a major role in this effort.
What is Community Guidelines:
YouTube implemented Community Guidelines to maintain a safe and vibrant community. So these guidelines are for making the creators aware about the do’s and don’ts of this platform. For example, they do not allow pornography, incitement to violence, harassment, and hate speech on this platform.
Results of Community Guidelines Enforcement July 2018 – September 2018:
According to YouTube – A channel is terminated if it gets three Community Guideline strikes in 90 days or has a single case of server abuse. Spamming is one of the main reasons for which channels are deleted. The statistics of removed channels given below(data was obtained from the Transparency Report of Google).
| REASON | % of removed channels |
| Child Safety | 4.5 |
| Nudity/Sexual | 12.6 |
| Spam, Misleading or spams | 79.6 |
| Multiple Policy Violation | 1.4 |
| Impersonation | 0.9 |
| Promotion of violence | 0.6 |
| Cyber Bullying | 0.3 |
| Abusive/Hateful | 0.1 |
YouTube relies on teams around the world to review flagged videos and remove videos which violate the terms of services of the company. For this task, they used Automated bots. These bots flagged 6,387,658 videos. Individual trusted Flagger(Flagged and removed 878,273 videos) and users (Flagged and removed 547,142 videos) also played an important role in flagging the inappropriate videos. Let’s have a look at the % of removed videos and their corresponding reasons:-
| REASON | % of removed videos |
| Spam, misleading and scams | 72.2 |
| Child Safety | 10.2 |
| Nudity or Sexual | 9.9 |
| Violent or Graphic | 3.4 |
| Others | 2.2 |
| Harmful | 0.8 |
| Cyber Bullying | 0.7 |
| Promotion of violence | 0.4 |
| Hateful or abusive | 0.2 |
Now let’s see what is the scenario of comments. Automated bots flagged 99.5% of removed comments. This indicates the intensive use of machine learning on YouTube recently. There were also 1066169 comments flagged by humans, which is far far less than the number of comments which were reported by bots.
