It's because their explanation is nonsensical on its face. How is a dislike "harassment"? How are genuine groups of people downvoting something an "attack"? YouTube isn't like HN where likes and dislikes are summed. People disliking a video doesn't change how many people liked it, both are visible.
This seems to come from a culture in which disliking someone, disagreeing with them or telling the they're wrong is the worst offense possible. In no way is it comparable to actual harassment, so of course people speculate about alternatives.
If someone seeks out content that they wouldn't ordinarily interact with just so that they can dislike/downvote it or post nasty comments - that's an attack.
If someone gets recommended a video and they don't like it, disliking is then constructive criticism.
But there are groups out there that actively look for particular types of content to dislike - without even watching it.
YouTube knows how someone arrived at a video - if they wanted to, they could only count likes and dislikes from people who had the content recommended to them. Or they could drop likes/dislikes from incomplete views. Actually I think they already do that, I recall reading something to that effect years ago.
It's all irrelevant anyway - their stated justification puts "harassment" first. Given the poisonous cancel culture that exists inside Google it's very obvious what their real motives are and the rest is thus all open to question. If it was 10 years ago Google's only justification for this sort of change would have been to improve the utility of the site, and we'd have believed it because that'd have been consistent with their other actions. In 2021? Not so much.
This seems to come from a culture in which disliking someone, disagreeing with them or telling the they're wrong is the worst offense possible. In no way is it comparable to actual harassment, so of course people speculate about alternatives.