Handling False Positives in Profanity Filters

One of the biggest challenges in automated content moderation is the false positive—when legitimate content gets incorrectly flagged as profane. The most famous example is the "Scunthorpe problem," named after the English town whose name contains an unfortunate substring.

Understanding the Scunthorpe Problem

The town of Scunthorpe has seen its name blocked by overzealous filters for decades. Other common false positives include:

• "Assassin" (contains a common profanity twice)
• "Cockpit" and "Peacock"
• "Shitake" mushrooms
• "Dickens" (the author)
• Many place names and surnames

These false positives frustrate users and can make your platform seem overly restrictive or poorly designed.

How Blasp Handles False Positives

Blasp uses several strategies to minimize false positives:

1. Word Boundary Detection

Rather than simple substring matching, Blasp considers word boundaries. This prevents "Scunthorpe" from being flagged while still catching actual profanity.

2. Context-Aware Processing

Our detection engine considers the context around potential matches to reduce false positives from compound words.

3. Curated Default List

Our default profanity list has been carefully curated to balance coverage with accuracy. Common false positives have been excluded.

Managing False Positives in Blasp

For paid users, Blasp lets you manage false positives directly from your dashboard. Navigate to API Settings and scroll to the Custom Word Lists section.

From there you can:

• Add custom false positives: Enter any word that should never be flagged in your API requests
• Toggle default false positives: Enable or disable words from our curated default list
• Remove entries: Delete any custom false positives you no longer need

Once you add "Scunthorpe" to your false positives list, it will never be flagged in your API requests. Changes take effect immediately.

Best Practices for Managing False Positives

1. Monitor user feedback: Users will tell you when legitimate content is being blocked
2. Review flagged content: Periodically audit what's being caught to identify patterns
3. Start conservative: It's easier to loosen restrictions than tighten them
4. Document your decisions: Keep track of why you added each false positive

Building a False Positive List

Consider creating false positive entries for:

• Place names relevant to your users
• Industry-specific terminology
• Common names and surnames
• Product names in your domain

The Balance

Perfect moderation is impossible. The goal is to find the right balance for your community—blocking enough to keep the environment safe while allowing genuine conversation to flow freely.

With Blasp's customization options, you can tune the filter to match your specific needs rather than relying on a one-size-fits-all solution.