Bot moderation vs human moderation: split by ambiguity

Bot moderation vs human moderation: split by ambiguity

The split isn't 'small community = human, large = bot'. It's: how ambiguous is the violation?

Bots handle unambiguous. Humans handle judgment.

Give the bot:
☐ Link/invite spam from new accounts.
☐ Banned keywords, slurs, obvious scams.
☐ Rate limits (X messages in Y seconds).
☐ First-post-must-be-text gates.

Keep for humans:
☐ Tone disputes and 'is this person a jerk or just blunt'.
☐ Off-topic that's borderline valuable.
☐ Anything involving a regular member.

The tradeoff: bots are fast and unemotional but produce false positives that feel cruel to good members. Humans read context but don't scale and burn out.

The SOP (standard operating procedure): bot acts on the clear 80%, flags the murky 20% to a human queue. Never let the bot ban a member with history — flag, don't act.

Save this for your next launch.