In Africa, where over a quarter of the world’s languages are spoken, many indigenous languages are underrepresented in AI development due to insufficient investment and data availability. Most AI tools, like ChatGPT, primarily train on English and major languages, leaving millions of speakers without access. To combat this, the Africa Next Voices project has created a significant dataset of 18 African languages, capturing 9,000 hours of speech across countries like Kenya, Nigeria, and South Africa. This initiative, supported by a $2.2 million Gates Foundation grant, aims to provide open-access data for developers to create AI tools that can translate and respond in local languages. Local farmers, like Kelebogile Mosime, have benefitted from agricultural apps that recognize indigenous languages, addressing challenges faced in farming. Experts emphasize that including African languages in AI is crucial for preserving culture and knowledge, ensuring communities aren’t left behind in the digital age.
Source link
Share
Read more