New benchmark evaluates π #AI detection tools across languages, π finding performance gaps π in low-resource languages and challenges β οΈ with distinguishing AI-translated and hybrid humanβAI text.
@jasonslucas1 @adaku_uchendu @penn_state @Visa