一作ACL 2026Findings· Findings of ACL
U buntu G uard: A Culturally-Grounded Policy Benchmark for Equitable AI Safety in A frican Languages.
Tassallah Abdullahi · Macton Mgonzo · Mardiyyah Oduwole · Paul Okewunmi · Abraham Toluwase Owodunni · Ritambhara Singh · Carsten Eickhoff
摘要
Current guardian models are predominantly Western-centric and optimized for high-resource languages, leaving low-resource African languages vulnerable to evolving harms, cross-lingual failures, and cultural misalignment. Moreover, most guardian models rely on rigid, predefined safety categories t...