languages of Africa
PulseAugur coverage of languages of Africa — every cluster mentioning languages of Africa across labs, papers, and developer communities, ranked by signal.
- 2026-08-10 funding Microsoft, Gates Foundation, and Google.org are funding 26 projects to develop AI for over 50 African languages. source
2 day(s) with sentiment data
-
Masakhane project boosts NLP research for African languages
The Masakhane project is a grassroots initiative focused on increasing the representation of African languages in Natural Language Processing (NLP) research. The organization aims to empower communities to shape and own…
-
New 'Model as a Library' architecture aims to improve AI for African languages
A new system architecture called Model as a Library (MaaL) has been proposed to address the challenges of deploying AI services for low-resource African languages. MaaL packages small, community-enrolled speech models a…
-
Tether releases open-source AI translation models for African and European languages
Tether has released open-source AI translation models specifically designed for African and European languages. This initiative aims to improve language accessibility and facilitate communication across diverse linguist…
-
Masakhane project champions NLP for African languages
The Masakhane project, a grassroots organization, aims to foster natural language processing (NLP) research in African languages. It highlights the severe underrepresentation of these languages in technology due to hist…
-
LLM representation geometry linked to language data availability
Researchers have identified that the performance disparity between low-resource and high-resource languages in large language models (LLMs) is linked to the geometric properties of their internal representations. A stud…
-
New SLMs Boost Machine Translation for 19 African Languages
A new research paper introduces TranslatePsy-AfriSLM, a suite of resources designed to improve machine translation for 19 Sub-Saharan African languages. The project includes curated parallel data, synthetic data tailore…
-
AI models fail African languages, losing 90% of safety signal
AI models demonstrate a significant degradation in safety alignment when tested with African languages, retaining less than 10% of the safety signal observed in English. This deficiency leaves speakers of these language…
-
Microsoft, Gates, Google fund AI for 50+ African languages
A new initiative aims to develop AI technologies for over 50 African languages, with funding from Microsoft, the Gates Foundation, and Google.org. The projects will focus on critical sectors such as healthcare, agricult…
-
New AfriNLLB models offer efficient translation for 15 African languages
Researchers have developed AfriNLLB, a suite of lightweight translation models designed for African languages. These models are derived from the NLLB-200 600M architecture, which has been compressed through layer prunin…
-
DONDO: Open ASR Models Released for 27 African Languages
Researchers have introduced DONDO, a collection of open-source automatic speech recognition (ASR) models specifically designed for African languages. These models are built upon the w2v-BERT 2.0 architecture and include…
-
Cross-lingual transfer learning shows mixed results for low-resource ASR
Researchers have explored cross-lingual transfer learning to improve automatic speech recognition (ASR) for low-resource languages. One study successfully used Sinhala to enhance Dhivehi ASR, achieving a 12.89% Word Err…
-
African languages face significant tokenization penalty in frontier LLMs
A new research paper reveals a significant "African Language Tax" in frontier large language models, where tokenizers assign substantially more subword tokens to African languages compared to English. This results in hi…
-
New benchmark tests LLM safety in African languages
Researchers have developed TukaBench, a new benchmark designed to evaluate the safety of large language models (LLMs) in seven African languages. This benchmark goes beyond simple translation by incorporating culturally…
-
Google releases WAXAL dataset for 27 African languages, AfriVoices-KE adds Kenyan languages · 2 sources tracked
Google Research has released WAXAL, a large-scale, open-access speech dataset covering 27 African languages, aiming to bridge the digital divide in speech technology. The dataset includes approximately 1,846 hours for A…