Abstract
Recent advances in artificial intelligence (AI) and natural languageprocessing (NLP) have improved the representation of underrepresentedlanguages. However, most languages, including Mali's 13 official nationallanguages, continue to be poorly supported or unsupported by automatictranslation and generative AI. This situation appears to have slightly improvedwith certain recent LLM releases. The study evaluated Claude AI's translationperformance on each of the 13 national languages of Mali. In addition to ChrF2and BLEU scores, human evaluators assessed translation accuracy, contextualconsistency, robustness to dialect variations, management of linguistic bias,adaptation to a limited corpus, and ease of understanding. The study found thatClaude AI performs robustly for languages with very modest language resourcesand, while unable to produce understandable and coherent texts for Malianlanguages with minimal resources, still manages to produce results whichdemonstrate the ability to mimic some elements of the language.