Research introducing FETCH!, a framework for systematically identifying coded expressions with dual meanings that evade content moderation systems.

• Identifies dog whistles by analyzing text usage patterns across different online communities
• Discovers expressions that appear innocuous to general audiences but carry hidden meanings for specific groups
• Provides innovative techniques for detecting emerging coded language without relying on pre-existing lexicons
• Demonstrates critical applications for enhancing content moderation and digital security measures

This research significantly advances online security by exposing disguised harmful content that traditional moderation tools miss, helping platforms better protect users while respecting free speech principles.

Making FETCH! Happen: Finding Emergent Dog Whistles Through Common Habitats