How to Reduce False Positives in Discord Moderation

False positives are the number one complaint about Discord moderation bots. They frustrate members, waste staff time, and erode trust in the moderation system. This guide explains how to reduce false positives using KenwoodBot AI-powered moderation and configurable settings.

Why false positives happen

Keyword-based filters trigger on exact matches without context. "I need to kill this process" triggers on "kill." "This game sucks" triggers on profanity. Regex helps but cannot distinguish context. AI moderation reduces false positives by analyzing full message intent — KenwoodBot users report up to 60% fewer false positives after switching.

Configure AI sensitivity

KenwoodBot lets you set sensitivity per channel and per action type. Start with "medium" sensitivity on general channels. Monitor the moderation log for a week. If you see too many false positives, reduce sensitivity. If genuine toxicity is slipping through, increase it. Each channel can have its own sensitivity level.

Use whitelists and exceptions

Create whitelist rules for common false positives. If "npm install" is being flagged, add it to the whitelist. KenwoodBot supports regex and exact-match whitelists. You can also whitelist specific channels where certain language is acceptable (e.g., off-topic channels).

Set up a review workflow

Use the KenwoodBot review queue where flagged messages are held for human review instead of deleted immediately. The AI auto-approves messages it is confident about and sends only uncertain cases to human moderators. This reduces staff workload while maintaining safety.

Monitor and iterate

Track your false positive rate in the KenwoodBot dashboard. A healthy rate is under 5%. Above 15% means configuration needs adjustment. Review the past 7 days of flags weekly and add exceptions as patterns emerge.

Frequently Asked Questions

What is a good false positive rate for moderation?

Under 5% is excellent. 5-15% is acceptable with tuning. Over 15% requires immediate configuration review and whitelist updates.

Does AI moderation eliminate false positives completely?

No moderation system achieves 100% accuracy. AI moderation reduces false positives significantly but cannot eliminate them entirely. Regular monitoring is still essential.

Conclusion

False positives are manageable with the right approach. Switch to AI-powered moderation, tune sensitivity per channel, maintain whitelists, and review your metrics weekly. Your members will thank you.