Without enforceability it'll go the same way as it has on Pixiv, the good actors will properly label their AI utilizing work, while the bad actors will continue to lie to try to maximize their audience until they get caught, then rinse and repeat. Kind of like crypto-scammers.
For context, Pixiv had to deal with a massive wave of AI content being dumped onto the site by wannabe artists basically right as the initial diffusion models became accessible. They responded by making 'AI-generated' a checkbox to go with the options to mark NSFW and adding an option for users to disable AI-generated content from being recommended to them. Then, after an incident of someone using their Patreon style service to pretend to be a popular artist, selling commissions generated by AI to copy the artist's style, they banned AI-generated content from being offered through that service.
I think that the idea is mostly to dictate culture. And I like the idea, not only for preventing fraud. Ever since the first Starship launches, the reality looks more incredible than the fiction. Go look up the SN-8 landing video, tell me that does not look generated. I just want to know what is real and what is generated, by AI or not.
I think that this policy is not perfect, but it is a step in the right direction.
I think that for now they're just going to use it as a means of figuring out what kind of AI-involved content people are ok with and what kind they react negatively to.
Personally, I've developed a strong aversion to content that is primarily done by AI with very little human effort on top. After how things went with Pixiv I've come to hold the belief that our societies don't help people develop 'cultural maturity'. People want the clout/respect of being a popular artist/creator, without having to go through the journey they all go through which leads to them becoming popular. It's like wanting to use the title of Doctor without putting in the effort to earn a doctorate, the difference just being that we do have a culture of thinking that it's bad to do that.
I think one of the bigger issues will be false positives. You'll do an upload, and youtube will take it down claiming that some element was AI generated. You can appeal, but it'll get automatically rejected. So you have to rework your video and figure out what it thought might be AI generated and re-upload.
For context, Pixiv had to deal with a massive wave of AI content being dumped onto the site by wannabe artists basically right as the initial diffusion models became accessible. They responded by making 'AI-generated' a checkbox to go with the options to mark NSFW and adding an option for users to disable AI-generated content from being recommended to them. Then, after an incident of someone using their Patreon style service to pretend to be a popular artist, selling commissions generated by AI to copy the artist's style, they banned AI-generated content from being offered through that service.