Reflection Essay: Evolution of language & context within harmful speech
Fulltext URI
Document type
Additional Information
Date
relationships.isAuthorOf
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Harmful speech does not have a stable lexical form: as platforms moderate explicit profanity and slurs, speakers shift toward indirect, coded, context-dependent expression, relocating harm from overt wording to implications, such as shared background knowledge. This reflection essay argues that surface text alone is unreliable evidence of harm in either direction: messages that appear hateful can be joking, ironic, affiliative, or counterspeech once context is known, while seemingly innocuous text or symbols can carry exclusionary or incitement force within a thread. The research agenda of this essay is to develop context-aware evaluation metrics and datasets and to integrate human moderators’ interpretive expertise into the design of next-generation content moderation systems.
