Streams containing informational or educational content that aim to share knowledge in a neutral, fact-based manner, rather than engaging in any kind of advocacy for an issue or candidate. For example, sharing the history of how votes in the US presidential election are counted to determine the next President, or merely encouraging individuals to vote or register to vote.
So saying for example Trump is a homophobic fascist should be allowed
Just because it's true doesn't mean it's not advocacy.
Propagandizing and "sharing knowledge in a neutral, fact-based manner" aren't mutually exclusive. The atomic unit of propaganda isn't lies, it's emphasis.
I have two facts that I intend to share in a neutral manner (and, for the case of this hypothetical, we will assume that “sharing knowledge in a 100% completely neutral, fact-based manner” is even possible).
I will call these Fact A and Fact B.
During the Super Bowl, I denote 30 seconds of airtime to Fact A, and denote only 5 seconds of airtime to Fact B.
Absolutely. Humanizing politicians is biased towards the status quo by distracting from the effects of their policies, which is literally the only relevance they have to our lives.
There's an implicit liberal, idealist bias in examining personal aspects of politicians instead of political economy and what factions in power selected that politician.
True neutrality, yes. But the average person sees neutrality as the appearance of neutrality, which is what propaganda revels in. It's why any both sides arguments are inherently propaganda on many topics, because just the very act of attempting to appear like there are two valid sides is in and of itself propaganda.
Climate change is a perfect example of this. Anthropogenic climate change is happening and even the oil companies are having to admit it publicly (after knowing about it for at least 60 years, but we've known this was an issue since 1890), but there are still tons of places who bring on denialists after yet another year of 'record breaking, once in a lifetime's storms.
As the op points put, it's going to be used as a reporting harassment. If it requires human intervention to decide, they might have bots or automatic actions based on number of reports.