So by that logic, if a tv station chooses to broadcast content that propagates racist misinformation and calls for violence, but the station only does it because that's what they algorithmically determine gives them the most viewership, that it's fine?

Your comparisons are not even close: Email and ISPs are not amplifiers of content. They do not curate messages.

Facebook's algorithms absolutely were responsible: they amplified the propaganda that led to deaths. And their money-first-research-second approach at censorship is part of the reason.

Rather than turning off the service, they let it run wild.

FB is to blame for aiding and abetting.

The problem with this analogy is that GMail and ISPs treat all traffic equally, where as Facebook and other "engagement"-driven social media use algorithms to prioritize content that generates the most engagement (which often happens to be the most outrageous, offensive or divisive content).

And yet they are all responsible for reporting CSAM found on their servers. These platforms are not dumb pipes, the ability is already there to decide what is allowed and what is not.