A new initiative is underway to improve the representation of African languages in artificial intelligence (AI) and other technologies. Researchers have noted a persistent lack of sufficient data for African languages in existing AI models, a problem this project is directly tackling.

The initiative, named Africa's Next Voice, has compiled a massive dataset of 9,000 hours of audio recordings from languages spoken in Kenya, Nigeria, and South Africa. This open-access data is designed to be a free resource for anyone looking to build language models, create speech-to-text tools, or develop translation services.

Of the more than 2,000 languages spoken across Africa, most have been left out of the technological revolution due to a severe shortage of available data. This new project, which will focus on 18 languages from the three partner countries, is a significant step toward bridging that digital divide.

The effort is backed by a substantial $2.2 million grant from the Gates Foundation. This information was obtained from the Ethiopian Artificial Intelligence Institute, as originally reported on the Nature website.