You don't need human eyes on every single tweet, just need enough people to review the flagged content. You could still have algorithms flag things and rank th

You don't need human eyes on every single tweet, just need enough people to review the flagged content. You could still have algorithms flag things and rank them by confidence levels. You could also hide the content first and queue it for human review instead of auto-banning the user. It's a lot of work, but I feel like there's a lot more that could be done.