One might evaluate the situation based on what I think is called a "preponderance of evidence", combined with an understanding that the legal system is both slow and tends towards innocence unless a crime is proven "beyond a shadow of a doubt".
A person may know how slow and different a legal decision is compared to what may be obvious and a reflection of reality, and therefore might arrive at a conclusion well before a system designed to be conclusive would.
The law is more about what can be proven than it is about what is true, and for people who know that, legal judgement stands separately from moral evaluation.
The whole philosophical backing of both "freedom of speech" and "innocent until proven guilty" is that the government doesn't itself have civil rights, only the rights explicitly outlined to it in the founding documents of that government (e.g. US Constitution).
Once you venture into private parties evaluating other private parties, you encounter a collision of rights. It's still freedom of speech and association to not want to do business with certain people, and as long as those certain people aren't of a protected class, this falls well within the moral concepts of both free speech and presumption of innocence.
In addition to the contracts, YT could easily say even the implication he may have done wrong is not good for their business to associate with. He needn't be convicted in a court of law for it to be bad business to continue to work with him.
Why would an ad need to run a script, use meaningful resources or even be from arbitrary domains?
Those are all practical arguments, but the point I'm making is that advertising doesn't have to be this way; it's the implementation that's the problem, not the theory.