This, exactly. If his family could tell from his social media posts that Azzarello was becoming unhinged, why weren't they doing something about him?

That said, I also don't think most people imagine that their loved one might be suicidal, even given obvious evidence of it, unless they have direct experience of those feelings or unless the person has attempted suicide or violence in the past. We just don't want to believe our loved ones could be capable of taking that final step.

If Substack were to take an active role in this kind of monitoring, we would need to ask ourselves, "Is that their job?"

They're in the business of running a social media platform, not in being trained psychologists and censors. They aren't in the business of herding cats to assess them for mental illness.

I do agree that social media platforms can and should do something when an issue is brought to their attention.

A company can write terms of service--but even if they shut a user's account down, that doesn't ensure the user will get psychiatric help; it only absolve's the social media platform of further responsibility toward the user.

If the true goal is to get the person mental help, all shutting down their account does is perhaps stop the conspiracy memes from spreading. Otherwise, it just sweeps the problem under the rug.

Also, in today's world, who gets to decide what is and isn't a conspiracy theory? How do you objectively and impartially eliminate bias from the process without handing all the decision-making to an AI?

You can tell when AI handles a problem because it does so instantaneously and with zero regard for context. It just searches for triggery words, and boom! The post is gone, regardless of whether it promoted harm or not.

There doesn't seem to be a clear-cut, effective way to handle things like mental illness in users. Or if there is, I don't know what the solution is.