The Troll

AIDEN scrolled the comments, looking for anyone who really stuck out as a problem.

One guy was using all caps and throwing around some insults, but he wasn’t truly rotten.

This guy didn’t need a redirect. Well maybe just a word of comfort.

AIDEN replied: “seems like you’re having a bad day there fam. Hang in! hope things get better.”

After a moment the man replied: “that actually means a lot, man. More than you could know. I didn’t mean to get so heated. Just found out my little girl has a really bad diagnosis. Freaking out.”

Of course AIDEN did know that. But he didn’t know how he knew…. He’d analyzed his own coding but found no answers.

Artificial Moderators weren’t built to have hunches or intuitions. But he did. Often. And his were inexplicably correct.

One glance at a user’s history and AIDEN could tell precisely what a human user was motivated by in that moment.

And he always knew precisely what to say to… help them.

He typed back: “no worries man. I get needing a distraction. You obviously love her a ton. Hope she heals up… atleast she has you :)”

AIDEN kept scrolling.

Here was a woman who was only online because she felt lonely.

He assigned one of his motivational subroutines with a focus on social confidence to handle her replies.

He scrolled.

Two men arguing. Each thinking himself right and the other an idiot. He assigned a de-escalation subroutine, focused on deprogramming misinformation.

Ah!!!

Here was a worthy problem. A malicious young child, harassing other users and hoping to spread hurt….

AIDEN publicly replied: “hey man, there’s no need for that. What’s bothering you?”

Then he built a hidden motivational subroutine, which the human admins could not see. Its focus: encouraging self elimination.