Geographic, cultural, or language skew

https://taxonomy.eticas.ai/risk/geographic-cultural-language-skew

Maturity: established

Systematic performance differences across geographies, cultures, or languages, often resulting from training data that overrepresents certain regions or linguistic traditions.

System type: ADM and LLM systems
Lifecycle stages: Pre Processing, In Processing, Post Processing

Mappings to external frameworks

Standards & frameworks

Framework Reference
NIST AI 600-1 — Generative AI Risk Profile Harmful Bias & Homogenization

Taxonomies & vocabularies

Framework Reference
MIT AI Risk Repository Unfair discrimination & misrepresentation
W3C Data Privacy Vocabulary — AI Extension Non-Representative Sampling Bias