Problem
Editors often rely on intuition to identify gendered or biased language, but research shows that terms like “female scientist”, “chairman”, or “manpower” persist in articles—even when they’re unnecessary or reinforce binary assumptions. While Wikipedia’s Manual of Style recommends gender-neutral phrasing, there's no lightweight tool to assist editors in spotting and reflecting on these patterns during the reading or drafting process.
Proposed Solution
Create a minimal browser extension (MVP) that passively scans the content of English Wikipedia articles for subtle gendered or outdated language patterns and flags them in a non-invasive UI.
Note: This is not a prescriptive tool. It offers context-aware prompts based on:
- Wikipedia Manual of Style – Gender-neutral language
- UN Gender-Inclusive Language Guidelines
- Widely used editorial references like the Chicago Manual of Style (non-binding)
Features (MVP Scope)
- On-page scanner activated by button or auto-load
- Flags common gendered terms (from curated list)
- Offers inline notes or side-panel explanations (“Why this was flagged”)
- Optional suggestions or questions to prompt reflection
- No writing or editing capability; this is a reader-facing tool
Stretch Goals
- Enable editors to submit new phrases or edge cases
- Expand to multilingual support (start with EN, flag DE/FR/ES for future)
- Respect quotation blocks, titles, and names via basic rule-based filtering
- Track frequency of flagged terms for analytics (client-side only)
Why?
Equity doesn’t start with new articles. It starts with how we describe people in the ones that already exist. This tool helps reduce linguistic bias without adding workflow bloat or triggering manual review drama. It’s designed for editors who care about tone, power, and representation—but don’t always have time to do a line-level bias audit.
Sample Terms to Audit
https://en.wikipedia.org/w/api.php?action=query&list=search&srsearch=%22unmanned%22&format=json
https://en.wikipedia.org/w/api.php?action=query&list=search&srsearch="female scientist"&format=json

