AI Farm Radio

Radio is a vital source of information for farmers in rural Africa.

In 41 countries across sub-Saharan Africa, Farm Radio International (FRI) reaches nearly 25 million listeners with its broadcasts that share weather forecasts and agricultural knowledge from experts, including CGIAR scientists, to help farmers make informed decisions. To ensure programming reflected the issues and experiences of farmers, FRI opened up a two-way communication channel so audiences could call in with questions and have them answered on air. Farmers jumped at the opportunity and the segment’s massive popularity quickly overwhelmed FRI, with only 1 out every 10 callers getting their questions answered in a timely manner.

To address this challenge, in 2021 FRI teamed up with CGIAR’s Digital Innovation Initiative to create an AI-powered automated tool that rapidly analyzes the content of audio messages spoken in a wide variety of local languages and dialects. The ability to quickly aggregate the questions and perspectives from farmers across the region will enable FRI producers to include millions of farmers in the conversation.

The results from the prototype are incredibly promising already the AI tool has achieved 97% accuracy when transcribing Swahili. The team is now building on this success by expanding the tool to include additional African languages, helping ensure more farmers can benefit from this technology.

Although the AI-powered tool remains in development, with more support it can be deployed at scale so farmers across Africa can be empowered with relevant information. In 2021 alone, nearly five million people improved their lives with information learned through FRI. By scaling this model to include other applications of AI-based audio analytics using local African languages, millions more people could change their lives too.

NEXT Climate-Smart Maize