Artificial intelligence is learning to speak the languages Nigerians use at home, in markets, schools and communities. But as companies build smarter tools using local speech, writing and cultural knowledge, an important question deserves more attention: who owns the knowledge that makes these systems work, and who benefits from it?
For a country with more than 500 languages, this is not just a technology debate. It concerns cultural preservation, economic opportunity, intellectual property and the rights of people whose everyday communication may become valuable training material for commercial AI systems.
Research published in the study NaijaNLP: A Survey of Nigerian Low-Resource Languages highlights the shortage of computational resources for Nigerian languages, including Hausa, Yoruba and Igbo. Building better language tools requires more high-quality data, but collecting that information responsibly is just as important as developing the technology itself.
Also Read: Nigeria Moves to Rethink Copyright Rules as AI Reshapes Creative Industry
How Nigerian Languages Are Powering AI Development
AI models learn patterns from large collections of text, speech recordings, translations and other information. For Nigerian languages, such material can come from books, news reports, online discussions, educational resources, recorded conversations and contributions from native speakers.
The challenge is that many Nigerian languages remain underrepresented in digital datasets. A study titled Does Generative AI Speak Nigerian-Pidgin? found that Nigerian Pidgin is poorly represented in generative AI, with systems relying on data that does not adequately capture its distinct linguistic characteristics.
This gap affects more than how naturally a chatbot responds. It can influence speech recognition, translation, educational applications and access to digital services. When an AI system misunderstands local expressions, cultural references or the meaning of a familiar phrase, its usefulness declines.
Initiatives such as Nigeria’s N-ATLAS project demonstrate the ambition to develop AI that better understands the country’s languages, dialects and accents. TechCabal’s report on the project describes the effort to develop a language model suited to Nigerian needs.
However, building such systems raises a fundamental issue: collecting language data is one thing, but determining who has the right to use it, and on what terms, is another.
The Ownership Question: Consent, Copyright and Community Rights
Not every word spoken in a community belongs to an individual in the same way a published book does. Yet recordings, written works, translations, original stories and organised datasets may carry legal rights or contractual restrictions.
Nigeria’s Copyright Act 2022 provides a legal framework for protecting eligible creative works. Whether particular material can be used to train an AI model depends on factors such as its nature, ownership, source, applicable exceptions and any relevant licence or agreement.
A 2024 study, Copyright in the Age of Artificial Intelligence, examines how Nigerian copyright law applies to AI training materials and text and data mining. The legal questions are significant because copying protected works into a training dataset may raise issues that cannot be settled simply by saying the information was publicly accessible online.
Language data also goes beyond copyright. A community’s oral history, traditional expressions or culturally sensitive knowledge may not fit neatly into conventional ownership arrangements. Speakers may contribute recordings without understanding how long the data will be stored, whether it will be shared or whether it could support a profitable commercial product.
Responsible AI development should therefore involve clear consent, transparent licensing, appropriate privacy safeguards and fair agreements with contributors. Where collective cultural knowledge is involved, developers should also consider community consultation and benefit-sharing arrangements.
Also Read: Why Nigeria’s Probe Into AI Companies Could Change Digital Journalism Forever
