The State of Large Language Models for African Languages: Progress and Challenges

2025-06-02

AI Summary

This study systematically evaluates the current support for approximately 2,000 low-resource African languages in state-of-the-art language models, revealing severe coverage gaps: only 42 languages are supported (<2% coverage), and only Latin, Arabic, and Ge’ez scripts are recognized—20 other active scripts remain entirely unsupported. Methodologically, we conduct the first comprehensive audit across six large language models (LLMs), eight small models, and six domain-specific models, performing multi-dimensional comparative analysis—quantifying language coverage, training data availability, script compatibility, and technical bottlenecks—along with root-cause diagnosis. Key contributions include: (1) identifying Amharic, Swahili, Afrikaans, and Malagasy as the only four consistently supported African languages; (2) cataloging 23 publicly available datasets; and (3) diagnosing four fundamental challenges—data scarcity, tokenization bias, prohibitive computational costs, and absence of standardized evaluation frameworks—thereby establishing a foundational benchmark and actionable roadmap for advancing AI fairness for African languages.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Knowledge Representation Languages

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Large Language Models (LLMs) are transforming Natural Language Processing (NLP), but their benefits are largely absent for Africa's 2,000 low-resource languages. This paper comparatively analyzes African language coverage across six LLMs, eight Small Language Models (SLMs), and six Specialized SLMs (SSLMs). The evaluation covers language coverage, training sets, technical limitations, script problems, and language modelling roadmaps. The work identifies 42 supported African languages and 23 available public data sets, and it shows a big gap where four languages (Amharic, Swahili, Afrikaans, and Malagasy) are always treated while there is over 98% of unsupported African languages. Moreover, the review shows that just Latin, Arabic, and Ge'ez scripts are identified while 20 active scripts are neglected. Some of the primary challenges are lack of data, tokenization biases, computational costs being very high, and evaluation issues. These issues demand language standardization, corpus development by the community, and effective adaptation methods for African languages.

Problem

• Evaluates African language coverage in LLMs, SLMs, and SSLMs
• Identifies lack of support for 98% of African languages
• Highlights challenges like data scarcity and script neglect

Innovation

• Comparative analysis of African language coverage
• Identifies 42 supported African languages
• Highlights lack of data and tokenization biases