I have little faith in crowd sourced human detection because I've been accused of botness a few times, but to the best of my knowledge I'm not one. Which is probably what a bot would say.
That's true, and I think it would need to be paired with some other features like having one account per person IRL, but it wouldn't be enforced unless the account is accused of being a bot. There are of course privacy concerns here, but it can only serve the purpose of verifying the account and does not have to be exposed to the public. I just think so far nothing creative or complex has been tried in this space.
