When Chika’s grandmother fell ill in their Enugu village, a critical barrier emerged. The doctor’s instructions, the medication labels, and the telehealth app were all in English, a language the 78-year-old barely understood. This scenario, repeated millions of times across Nigeria in healthcare, education, and finance, underscores a profound digital exclusion. The global AI revolution has been built on a foundation of English, Mandarin, and European languages, largely ignoring the over 500 native languages spoken in Nigeria. Now, a determined coalition of linguists, tech entrepreneurs, and researchers is working to change that, building AI that can truly hear and speak in our mother tongues.
This mission, known as “low-resource language NLP,” is one of the most urgent and complex challenges in African tech. Natural Language Processing (NLP), the branch of AI that enables machines to understand text and speech, requires vast amounts of written data to learn. For languages like Yoruba, Hausa, and Igbo with rich oral traditions but relatively smaller digital footprints, this creates a “data famine.” The first breakthrough is happening in data collection. Initiatives like Lanfrica and Masakhane, a pan-African research collective, are crowdsourcing thousands of translated texts, transcribing oral histories, and digitizing local radio broadcasts to build the first high-quality datasets for indigenous languages.
The practical applications are transformative. Startups are creating voice-activated assistants that respond to Pidgin English. Vatebra, a tech firm in Lagos, developed a prototype for a “Drivern” assistant that understands Nigerian traffic commands and slang. In education, Kọ́ńfíḿ is piloting an app that uses speech recognition to help children learn to read in Yoruba, providing real-time pronunciation feedback. For the legal and civic sphere, JusticeText is exploring AI tools to transcribe and translate court proceedings from local languages into English, promoting transparency and accessibility.
Perhaps the most critical application is in healthcare communication. Researchers at the African Centre for Technology (ACENT) are fine-tuning open-source AI models to power SMS-based health bots that can conduct basic medical triage in Hausa and Kanuri, bridging the gap in rural areas where community health workers are overwhelmed.
The effort extends beyond translation to preservation. As younger generations urbanise, fluency in native languages declines. AI tools like those from Ethnologue are being used to analyse speech patterns and document linguistic structures, creating living digital archives. “We are not just building tools; we are building arks,” says Feyisayo Anjorin, a computational linguist at the University of Ibadan. “We are encoding our proverbs, our humour, our cultural nuances into silicon so they survive for the next generation.”
The challenges are immense. Beyond data scarcity, there are technical hurdles. The tonal nature of many Nigerian languages (where a word’s meaning changes with pitch) confounds standard speech algorithms. Furthermore, dialects vary widely; the Hausa spoken in Sokoto differs from that in Maradi, Niger. Building one monolithic “Hausa AI” may not be sufficient. There’s also the risk of “digital dialect hegemony,” where an AI model trained primarily on one dialect marginalizes others.
Funding is another obstacle. Most global AI investment targets profitable, large-language markets. Pioneering Nigerian NLP work often relies on grants, academic passion, and volunteer efforts. Advocates are pushing for the National Language Policy to have a digital component, mandating public datasets in major Nigerian languages and creating incentives for private sector development.
The promise, however, is a more inclusive digital future. Imagine a farmer in Katsina getting voice-based market updates from an AI in flawless Hausa. Envision a student in Ilorin researching online using a search engine that understands Yoruba queries. This is the vision driving the community.
As Ada Nduka Oyom, founder of She Code Africa, states, “True technological sovereignty doesn’t mean just using tools built elsewhere. It means building the tools that understand the very soul of our communication. When an AI can grasp the proverbial wisdom in an Igbo ‘ili,’ that’s when we’ll know the digital age has truly arrived in Africa.”
The journey to teach AI our languages is more than a technical project. It is an act of cultural affirmation and democratic necessity, ensuring that the Fourth Industrial Revolution does not leave the majority of Nigerians behind, silent in a world of talking machines.
- business a.m. commits to publishing a diversity of views, opinions and comments. It, therefore, welcomes your reaction to this and any of our articles via email: comment@businessamlive.com
Olusoji Adeyemo is a professional with over 17 years of experience. Currently serving as an Azure Application Innovation & AI Specialist at Microsoft UK, he has held key positions at Wipro, Huawei Technologies, Oracle, and Dell, showcasing his expertise in cloud infrastructure, Application modernization, and Business continuity solutions. He holds a Master’s degree in Computer Science with distinction from the University of Hertfordshire and Caleb University. He is currently running his PhD research in Explainable AI and ML. He is also certified in various cloud and project management technologies, including Microsoft Azure Expert, Google Expert, AWS and Scrum. He can be reached at mastersoji@gmail.com and on Linkedin: https://www.linkedin.com/in/olusoji-adeyemo/
Africa’s energy wealth: Why good governance must power a just transition