It shouldn't be on the ISPs, it should be on the SERVICES that USE the ISPs.
I'll give you a perfect example, the Uvalde shooter.
He had been using a French social media platform called Yubo where he posted animal abuse videos and threatened to rape and murder other users.
He was reported to Yubo, REPEATEDLY, and Yubo did nothing.
Maybe we need to make social media companies mandatory reporters in cases like this? Rather than just ban a user wholesale, increase monitoring of them and report the account to local authorities?