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 ASR and over 565 hours for TTS, collected through collaborative efforts with African academic and community organizations. Concurrently, a new dataset called AfriVoices-KE has been published, featuring around 3,000 hours of audio across five Kenyan languages, with a mix of scripted and spontaneous speech. Both initiatives aim to foster the development of inclusive voice-enabled technologies and preserve linguistic heritage. AI
IMPACT These datasets are foundational for developing inclusive speech technologies and preserving linguistic diversity in underrepresented regions.
RANK_REASON The cluster describes the release of large-scale speech datasets for African languages, which constitutes a research milestone in AI.
Read on Google AI / Research →
- Abdoulaye Diack
- Creative Commons license (CC-BY-4.0)
- Google Research
- Sub-Saharan Africa
- Tavonga Siyavora
- WAXAL
- AfriVoices-KE
- Kenyan languages
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
