ISPs and telco networks are on a different layer (physical, and transport layers) of the stack than social media (application layer). ISPs and telco networks can and do perform traffic routing shaping, and throttle or cut off abusive users who consume too much bandwidth, or run a high-traffic web server from their home network. Because these actions affect other users of their networks. But if someone uses Comcast to post a sweary rant on the interwebs - it makes no difference to other Comcast customers. So they don't (at present, though the repealing of net neutrality now allows it) and shouldn't moderate actions at the application layer.

Social media is the opposite. Abusive users of those networks operate at the application layer, and can spoil the experience of other users at the application layer but (likely) not at the network layer. So they moderate user activity at the application layer.

In each case it's about trying to ensure bad actors don't ruin other customers'/users' experience. It's just done in different ways depending on what part of the network stack the bad actors do their work in.