The thing is, Twitter has data on how people responded to those two Tweets. If Trump gets prosecuted, I expect we will see responses to Trumps tweets as evidence in court. My guess is they can make a statistical causal claim. Example, what if some of those who broke into the capital earlier replied to one of Trump's tweets in such a way as to suggest they were incited by Trump?
Were those actual people, or bots? Did the people say "yes, let's do a violent thing" or did Twitter interpret "yes, we will not give up" as a response to a dog whistle? There is no analysis showing how they came to that conclusion, and there is no trust in them from those who affected by these bans.
