Something doesn't have to be "officially" illegal for laws to suppress it. If the law gives no penalty for a false-positive block, but very high penalties for any false-negative failure to censor, then companies will be incentivized to err on the side of censoring legal speech.
I don't disagree about the risk of chilling effects, but the methodology of this report is so embarrassingly, deceptively bad that I don't think any conclusion can be drawn from it. They basically snapshot a comment section, then return repeatedly, and enumerate the comments that disappeared. Without any insight into who (ie the platform, the poster, or the channel/page owner) removed the comment and why, this is like performing a legal analysis of the leaves that disappeared from lawns in autumn.
Yes, that's the primary complaint of the article, but the actual report just doesn't back that up. There's no mention of false positives or negatives. Probably because, as the report notes, it doesn't know who removed the comment, let alone what the rationale was. So it could be that they are attempting to comply with broad and vague laws, or they could just be against hosting MLP memes.