The problem of vetting YouTube videos isn't necessarily something that can be solved by an algorithm, at the end of the day. Lots of the simple cases, sure - but such an algorithm won't work on even the most moderately challenging content. After all, moderating complex content such as video doesn't just require a well-trained neural network to recognise objects portrayed in images and sounds in audio... deciding whether a video is appropriate or not for monetisation is something that requires an explicit understanding of the social and cultural context which the video is going to be viewed in - which is both highly subjective and difficult to express in an algorithm.
NB;
- Neural networks (and other AI techniques) can be tricked and manipulated (see [0,1]; examples of adversarial images)
- Machine learning algorithms can induce or confirm bias through inadequate training data and poor assumptions. (see [2]; a book about bias in machine learning algorithms)
[0]: https://arxiv.org/abs/1510.05328
[1]: http://www.evolvingai.org/files/DNNsEasilyFooled_cvpr15.pdf