I wonder why twitter will not use Machine Learning to remove terrorism propaganda from their social network? I mean you can train neural networks to recognize terrorism spreading accounts and tweets. It's not black magic, NLP to get intent + classification of known/blocked terrorist accounts and just train. Then show message "Sorry but this tweet looks like abusing our rules, please wait until our staff will check it" or something like that. If you do nothing where you can then you are just accomplice to spreading terrorism.