Small-scale innovation that can automate and educate with wit and knowledge.

Seems noble, but it's only a matter of time before bots auto-respond to things they don't like.  In fact, I think I've seen some of that already with the Caitlyn Jenner thing.   An automated bot detected masuline vs feminine pronouns and sent nastygrams to the author.  

It's probably best just to block all bots on Twitter, Facebook, Instagram, LinkedIn, wherever.

On Reddit some bots are helpful in auto-linking to useful sources on Wikipedia or the Dictionary.

Reddit lets those bots live. Bad bots get killed quickly.

I think Twitter has a similar ecosystem where the good bots can help.

But you're right that they have to stay useful and witty. Without that people will block them.