Artificial intelligence-powered tools are created to make our everyday lives easier. However, these tools are primarily designed for languages like English, French, and Spanish, leaving billions of people who speak other languages without access. Researchers in Africa are determined to change this. In a recent study published in the journal Patterns, scientists have outlined a plan to develop AI-driven tools specifically for African languages.
Developing AI Tools for African Languages
Kathleen Siminyu, an AI researcher at the Masakhane Research Foundation, believes it’s essential to create AI tools for African languages. She emphasizes the importance of inclusion and representation in advancing language technology.
To develop AI tools for African languages, researchers rely on a branch of AI called natural language processing. This technology trains computers to understand human languages. However, when there is limited data available in a specific language, such as African languages, the effectiveness of these tools diminishes. To bridge this gap, the research team identified key players who are involved in developing African language tools. These individuals include content creators, linguists, software engineers, and entrepreneurs who establish the necessary infrastructure.
Four Themes for African Language Tool Design
Through interviews with these key players, the research team identified four essential themes to consider when designing AI tools for African languages:
- Africa’s multilingual society: Africa’s multitude of languages holds significant cultural importance and plays a crucial role in education, politics, and the economy.
- Supporting content creation: Basic tools like dictionaries, spell checkers, and keyboards for African languages are necessary, along with removing financial and administrative barriers for translating government communications.
- Collaboration between linguistics and computer science: Fostering partnerships between linguists and computer scientists is crucial for creating effective African language tools. The focus should be on building human-centered tools that empower individuals.
- Ethical data practices: Developers must prioritize communities and ethical practices when collecting, curating, and using data.
Siminyu believes this study will help coordinate efforts in building impactful language tools for African languages. By highlighting the priorities and investments necessary, this research will guide the development of various language tools, including translation services and content moderators.
Expanding the Study and Future Goals
The research team plans to expand the study by including more participants to better understand the communities that AI language technologies will impact. They also aim to identify and address barriers that may hinder people’s access to these technologies. Additionally, the team hopes their findings will contribute to the preservation of indigenous African languages.
Siminyu envisions a future where Africans have the same quality of life, access to information, and opportunities as those who speak widely supported languages like English, French, and Mandarin.