Your logic makes sense but is not how these moderation services actually work. When I used my own phone number to create a Twitter, I was immediately banned. So instead I purchased an account from a service with no issues. It’s become impossible for me at least to use large platforms without assistance from an expert who runs bot farms to build accounts that navigate the secret rules that govern bans.

Spam is a behavior, not a fundamental trait of the actor.

Would be interesting to make a service where spammers have to do recaptcha-like spam flagging to get their account unlocked.

Which definition of spam are you operating under? I think it is a fundamental trait of the actor.

So you would expect a spammer to only ever post spam, even on their own personal account? Or a spam emailer to never send a personal email?

I know sales bros who live their live by their ABCs - always be closing, but that's besides the point. if the person behind the spam bot one day wakes up and decides to do turn over a new leaf and something else with their life, they're not going to use the [email protected] email address they use for sending spam as the basis for their new persona. thus sending spam is inherit to the [email protected] identity that we see - of course there's a human being behind it, but as we'll never know them in ant other context, that is who they are to us.

> But you likely aren't, and most people likely aren't either.

Is there any evidence of this? 1% bad content can mean that 1% of your users are bad actors, or it can mean that 100% of your users are bad actors 1% of the time (or anything in between.)

I assume all of us have evidence of this in our daily lives.

Even the best people we know have bad days. But you have probably also encountered people in your life who have consistent patterns of being selfish, destructive, toxic, or harmful.

> you have probably also encountered people in your life who have consistent patterns of being selfish, destructive, toxic, or harmful.

This is not evidence that most bad acts are done by bad people. This is evidence that I've met people who've annoyed or harmed me at one or more points, and projected my personal annoyance into my fantasies of their internal states or of their essence. Their "badness" could literally have only consisted of the things that bothered me, and during the remaining 80% of the time (that I wasn't concerned with) they were tutoring orphans in math.

Somebody who is "bad" 100% of the time on twitter could be bad 0% of the time off twitter, and vice-versa. Other people's personalities aren't reactions to our values and feelings; they're as complex as you are.

As the OP says: our definitions of "badness" in this context are of commercial badness. Are they annoying our profitable users?

edit: and to add a bit - if you have a diverse userbase, you should expect them to annoy each other at a pretty high rate with absolutely no malice.

I think the point is that anyone and/or everyone can be a bad actor in the right circumstances, and moderations job is to prevent those circumstances.

We have laws around mobs and peaceful protest for a reason. Even the best people can become irrational as a group. The groupmind is what we need controls for: not good and bad people.

> and spaces that allow bad actors to grow

I believe that's GP's point! Any of us has the potential to be the bad actor in some discussion that gets us irrationally worked up. Maybe that chance is low for you or I, but it's never totally zero.

And even if the chance is zero for you or I specifically, there's no way for the site operators to a priori know that fact or to be able to predict which users will suddenly become bad actors and which discussions will trigger it.