I don’t know honestly. Really, with AI it would be pretty difficult to be foolproof. I’m thinking of the MIT card counting group and how they played as archetypal players to obscure their activities. You could easily make an account that upvoted content in a way that looked plausible. I’m sure there are many real humans that upvote stories positive to one political party and downvote a different political party. Edit: I mean fuck, if you wanted to, you could create an instance just to train your model. Edit 2: For that matter, you could create an instance to bypass any screening for botters…
This is like a $30 sandwich from one of the most famous delis in the country. I would not describe this as a typical American sandwich.