The Rise of Multilingual AI in Africa: Language Models, Inclusion, and Local Context

GSMA and Pleias Launch CommonLingua to Support 61 African Languages

SMA and Pleias have launched CommonLingua, an open-source language identification model designed to strengthen multilingual AI development across Africa. The initiative supports 61 African languages and aims to improve how artificial intelligence systems recognize and process African digital content, which has historically been underrepresented in global AI platforms.

The project seeks to provide developers with accessible AI infrastructure that can power more inclusive digital tools and applications. By addressing the shortage of African language datasets, CommonLingua represents an important step toward ensuring African users can participate more fully in the evolving AI economy and digital ecosystem.

Source: GSMA

Cohere Launches Open Multilingual AI Models to Expand Global Language Inclusion

AI company Cohere has introduced a new family of open multilingual language models focused on improving AI performance across a broader range of global languages, including many low-resource languages often excluded from mainstream AI systems. The launch reflects the growing industry focus on building more inclusive artificial intelligence technologies that can serve diverse linguistic communities worldwide.

The models are designed to improve accessibility and create more culturally relevant digital experiences for users in emerging markets. Experts believe that advances in multilingual AI could accelerate AI adoption across Africa and other regions where language diversity remains a major barrier to digital inclusion and technological participation.

Source: TechCrunch

UNESCO Warns African Languages Risk Being Left Behind in AI Development

UNESCO has raised concerns that many African languages remain significantly underrepresented in global AI systems, warning that this imbalance could contribute to digital exclusion and cultural inequality in the future of artificial intelligence. The organization says stronger investment in multilingual datasets, inclusive AI education, and local innovation ecosystems is essential to support equitable technological growth.

According to UNESCO, most major AI systems continue to prioritize dominant international languages, leaving many African communities at risk of being excluded from future digital tools and services. Experts warn that failing to address language inequality in AI may deepen existing digital divides and limit opportunities for inclusive innovation across the continent.

Source: UNESCO

Microsoft Research Launches Paza to Advance AI Speech Recognition for African Languages

Microsoft Research has introduced Paza, a new initiative focused on improving automatic speech recognition technologies for low-resource languages across Africa. The project includes PazaBench, the first ASR leaderboard dedicated to evaluating speech recognition systems for African languages, supporting 39 languages and benchmarking more than 50 speech models.

Microsoft also developed speech recognition models for six Kenyan languages, including Swahili, Kikuyu, Maasai, Somali, Dholuo, and Kalenjin. The initiative emphasizes real-world usability in low-bandwidth and rural environments, with early testing conducted directly with farmers using mobile devices in local communities. Researchers say the project aims to reduce digital exclusion by improving AI accessibility for African language speakers.

 Stay Ahead in Africa’s AI Economy!

Subscribe to AI@DCADigest on Substack for insights on multilingual AI, digital inclusion, emerging technologies, and the future of innovation across Africa.