All About The African Company Using AI To Promote Local Languages

All about the African company using AI to promote local languages

Amid emerging Artificial Intelligence (AI) technologies being developed globally to aid language processing, text, and writing, an African company has risen to the challenge of introducing application software focused on local African languages.

The company, Ghana NLP, began operations in 2019 introducing application software that focuses on machine translation, automatic speech recognition, and text-to-speech. It has applied this to languages in three African countries, Kenya, Ghana, and Nigeria.

According to the Director of Products of the company, Lawrence Adu-Gyamfi, the initiative was introduced to highlight the gaps in local language literacy in many African countries as well as promote the learning and use of these languages by citizens of the respective countries.

The ‘Khaya’, which is its flagship app combines three natural language processes through machine translation, speech recognition, and text-to-speech to facilitate translation between English and African local languages.

“Language is a huge part of our identity and it’s something that needs to be hugely preserved and we cannot lose this. That’s why it’s important that we bring attention to this and keep developing tools that will bring attention to it and keep on developing tools that will keep them in existence, even if it’s for a language that is spoken by 100,000 people, it’s not too small to forget it.

“The features of Natural language processing are in this software, machine translation, speech recognition, and text-to-speech. So basically what we are trying to do is to say I want to be able to translate between English and any local language but I have it in text and I want to see the corresponding text in the other language but you know that computers don’t understand our text so you are trying to find a way to find a mapping between a text in one language to another language and so what exists between that mapping is what we call our model or a machine learning model,” explained Lawrence.

The app which is being used by students, individuals, and companies mainly focused on data and the agricultural sector has so far benefitted about 100,000 persons who have downloaded it via Android and about 20, 000 others who use it on Apple IOS.

So far, the Khaya app processes three Kenyan languages, Kikuyu, Kimeru, and Luo. It also processes the Yuroba language in Nigeria and six other languages in Ghana, Twi, Fante, Gurune, Ewe, Dagbani, and Ga.

Lawrence Adu-Gyamfi further noted that there are plans underway to further expand their research on generative AI, to develop tools that will be more relatable and applicable to more fields including the medical field and the agricultural sectors to assist farmers and other individuals who may not be English oriented.

“We are trying to use the voice where people will use it and hear it. We want to use our tools and apply them in the medical field, and the agricultural sector, to build tools that will actually mean something,” he told GSW’s Wonder Hagan.