500 Nigerian Languages Face Extinction If They Are Not AI Compliant

There are over 500 languages spoken in Nigeria and all of them would be dead in 50 years if they are not digitized and made compliant with artificial intelligence. This was stated by a linguistics professor and Head of the Linguistics Department at the University of Ibadan, Dr Demola Lewis.

According to Dr Lewis while delivering a welcome address at the UNESCO/Infinix STEM Education Project in Ibadan, if the large language models used to train artificial intelligence do not speak these Nigerian and African languages, then they will be dead heading into the future.

“One of the greatest resources that this country has is language. And believe it or not, we may say that we have 500 plus languages but they all will be dead if they are not AI-compliant in the next 50 years. So if large language models are not speaking our languages, if it has not get Yoruba and Igbo and Hausa in it, we are talking about dead languages in the IT world,” he said.

                                                                                                                                                                                                                   Dr Demola Lewis

Nigerian languages need to be integrated into language models
Big tech organisations worldwide have continued to develop their artificial intelligence tools. Probably the most powerful of them is the large language model (LLM) which is a computational model that processes languages. This tool has to undergo a process of deep machine learning to effectively consume vast amounts of text and learn statistical relationships that would enable it to use the language efficiently.

The LLMs, however, require huge materials, texts, and perhaps context from other languages to effectively use the language. This is how, apart from English which has become a first language for search engines like Google, others like Igbo, Hausa, Yoruba and Swahili have been included with Google poised to extend the options for other languages within Nigeria and Africa at large.

And for Dr Lewis, any language that is not digitized and used by the computer is well on its way to becoming irrelevant in the future. He pointed out that language trainers are not just training humans, but are also training a computer. This means there has to be deep learning by the computer otherwise, they will have no use in the future.

He noted that things are changing very quickly and individuals and organisations understand that they must be AI-compliant or find themselves grappling with relevance in their respective fields. He also said that big tech companies like Google are deploying experts in different languages to put them in a format that is machine-readable. This means they have to annotate the languages so that the machine can begin to learn the language and understand how the words are used.

“Except our languages are so resourcefully put into these systems, we are not going to have them. We are going to find ourselves in a situation where the next generation of children are going to be out of touch with African languages and culture because the language goes with its culture. So you are going to have a unison world where everybody thinks and speaks the same way and does everything the same way and that is not going to be for Africa. So it is important that we emphasise the fact that they have to be digitized,” he said.