Experts seek to close the gap in the language

Although there is a large proportion of languages ​​in Africa – more than a quarter of some estimates – many are missing when it comes to developing II.

This is also the question of lack of investment and easily available data.

Most AI tools, such as Chat GPT used today, are trained in both English and other European and Chinese languages.

They have a large number of text on the Internet from which you can turn.

But since many African languages ​​are mostly spoken, not recorded, there is a lack of text to teach the II to make it useful for the speakers of these languages.

For millions on the continent, this means that remains over.

Researchers who have tried to resolve this issue recently released what is considered the largest set of African data.

“We think that in our own languages ​​you dream of them and interpret the world through them. If the technology does not reflect it, the whole group risks,” – says BBC University of Pretoria Vukosi Marivate, who worked on the project.

“We are experiencing this revolution in the II, imagining that you can do everything you can do. Now imagine that there is a part of the population that simply does not have such access because all the information in English.”

The next project of Africa gathered linguists and computer scientists to create data sets ready for AI in 18 African languages.

It may just be a small part of more than 2000 languages ​​that are estimated to speak on the continent, but those who participate in the project say they hope to expand in the future.

For two years, the team recorded 9,000 hours of speech in Kenya, Nigeria and South Africa, fixing everyday scenarios in the field of agriculture, health and education.

The recorded languages ​​included Kikuyu and Dhluo in Kenya, Husse and Yoruba in Nigeria, and Iszul and Tivan in South Africa, some of whom say millions.

“You need some basis to start with it, and this is what is in Africa of the following votes, and then people will build it and add their own innovations,” says Professor Marivate, who has led research in South Africa.

His Kenyan colleague, computing linguist Lillian Vansare, says that the recording of speech on the continent meant the creation of data aimed at reflecting how people really live and talk.

“We have gathered votes from different regions, centuries and layers, so it is as inclusive as possible. Great technology can not always see these nuances,” she says.

The project was made possible by the Gates 2.2 million (1.6 million pounds.

The data will be open access, which allows developers to create tools that translate, rewrite and respond to African languages.

According to prof. Mariivate, already there are small examples of how indigenous languages ​​used in II can be used to solve real -life problems.

The Kelebogile Mosime farmer runs 21 hectares in Rustustburg, the heart of the South Africa platinum.

The 45-year-old guy works with a small team of vegetables, including beans, spinach, cauliflower and tomatoes.

It started only three years ago, from cabbage, and help it use an app called AI-Farmer, which recognizes several South African languages, including Cassot, Isisul and Afrikaans to help solve different problems.

“How anyone else learns to engage in the farm, you face a lot of problems,” says Ms. Masim.

“Daily I see the benefits of using your home language STSWANA in the app when I face problems on the farm, I ask for anything and get a useful answer.

“For someone in the countryside, as I do not expose technology, it is useful. I can ask about different options for control of insects, it is also useful for the diagnosis of diseased plants,” it is widely enchanted.

Lelapa AI is a young South African company that builds AI tools in African languages ​​for banks and telecommunications firms.

For its Pelonomi Moiloa, what is currently available is very restrictive.

“English is the language of opportunities. For many South Africans who do not speak this, it is not just uncomfortable – it can mean that the lack of basic services such as health care, banking and even state support,” she says the BBC.

“The language can be a huge barrier. We say that it should not be.”

But this is more than doing business and convenience.

For Professor Marivate also there is a danger that without African language initiatives can be lost

“Language is access to fantasy,” he says.

“Not only words are a story, a culture, a knowledge. If indigenous languages ​​are not included, we lose more than data; we lose ways of vision and understanding the world.”