Remember: During the shutdown, YouTube replaced the moderators with artificial intelligence, which is responsible for detecting and managing all sensitive content published on the platform. Today, Youtube goes back and calls back its good old people.
But what motivated this backtracking? In fact, the algorithm developed by YouTube multiplied the overkill and was far too aggressive. As a result, more and more content was removed, sometimes without good reason, which upsets users. YouTube has therefore decided to return to a more humane moderation by recalling its moderators in the flesh.
Twice as many erasures
When artificial intelligence was in power, millions of videos were deleted. 1
For YouTube, it’s way too much. The platform also admitted that its algorithm was far too incredible. New reverse platform that in March completely assumed the difficulty level of its AI while knowing it could make mistakes.
Back to the office
But today, YouTube employees (including moderators) are gradually returning to the office after locking up. These employees can therefore resume their moderation tasks. In the end, we can clearly speak of a recognition of weakness from YouTube and this whole story has made it possible for everyone to discover flaws and weaknesses in such algorithms.
Yes, AI always has trouble detecting certain subtleties in language as well as racist or conspiratorial comments. To discover them, nothing like real people. In addition, AI sometimes acted strangely like the time it removed critical comments from the Chinese Communist Party during a video. In short, not everything is quite clear.
The only real strength of the YouTube algorithm: speed. According to YouTube: “over 50% of these 11 million videos were deleted without a single view by a real user, and over 80% were deleted with less than 10 views”.