Unlocking Nigeria’s AI Potential: The Launch and Future of N-ATLAS for Indigenous Languages

Every Thursday, Delve into AI brings you in-depth perspectives on Africa’s evolving artificial intelligence landscape. This series explores the intersection of AI with culture, governance, commerce, and society, spotlighting the innovators and initiatives shaping the continent’s AI future.

Introducing N-ATLAS: Nigeria’s Indigenous Language AI Model

In a strategic move to assert leadership in ethical and inclusive AI development, the Nigerian government unveiled N-ATLAS in April 2024. Developed in partnership with Lagos-based startup Awarri, this AI model is designed to understand and process Nigeria’s diverse local languages and accents. Currently, N-ATLAS is accessible at no cost for research, prototyping, and small-scale applications. However, commercial use involving over 1,000 users requires a separate licensing agreement.

This initiative marks a significant step toward Nigeria’s ambition to secure AI sovereignty on the global stage. Yet, the true test lies in whether local innovators, policymakers, and entrepreneurs can transform this foundational technology into practical solutions for education, agriculture, healthcare, and beyond-despite ongoing challenges like limited funding and infrastructural constraints.

Building Blocks for Nigeria’s AI Ecosystem

For the average Nigerian, N-ATLAS may not yet be a consumer-facing product. Instead, it serves as a critical infrastructure layer aimed primarily at developers and researchers. Hosted on Hugging Face, a popular open-source platform, the model is currently tailored for technical users rather than everyday consumers.

Sunday Afariogun, lead project engineer at Awarri, emphasizes that the initial release is developer-centric. “While the public might celebrate the announcement, many won’t immediately know how to leverage it,” she explains.

By providing a base model that comprehends Nigerian dialects and accents, N-ATLAS fills a crucial gap. Bilesanmi Faruk, CTO and co-founder of Lagos-based AI startup Lena, highlights the limitations of existing global AI models, which often lack cultural and linguistic relevance for Nigerian users. “We spend excessive time adapting foreign models or end up with tools that don’t resonate locally,” he notes.

Faruk is already integrating N-ATLAS into Lena’s offline educational app, enabling children in rural areas to learn in their native tongues while receiving real-time feedback. “This technology unlocks access to quality education for over 100 million learners,” he says.

Similarly, NLP researcher Zainab Taairu underscores the value of local open-source models in healthcare projects. “Previously, finding reliable datasets for African voices was a major hurdle. N-ATLAS simplifies building solutions grounded in authentic local contexts,” she shares.

Though still behind the scenes, N-ATLAS has inspired ideas for sector-specific applications-from farmers consulting hotlines in Igbo about crop diseases to patients describing symptoms in Nigerian-accented English for AI-assisted diagnosis.

Joshua Firima, co-founder of voice and text AI startup KrosAI, envisions a future where AI-powered phone systems seamlessly serve Nigerians in their native languages. “True success is when AI becomes an invisible, everyday tool for millions,” he asserts. However, Faruk stresses the need for robust documentation, community engagement, and developer support to maximize the model’s impact.

Challenges on the Path to AI Empowerment

While N-ATLAS is a promising milestone, it also highlights persistent obstacles in Nigeria’s AI landscape-namely, data scarcity, limited computing infrastructure, and insufficient funding.

Afariogun warns against postponing AI development until infrastructure catches up. “Waiting would only widen the gap,” she says. Although Nigeria’s data centers are improving, only a handful can support the GPU-intensive workloads required for training advanced AI models.

Currently, Awarri depends on international cloud providers like AWS and Google Cloud, which offer reliability but raise concerns about data sovereignty and control.

Faruk points out that the high costs of GPU resources pose a significant barrier for small teams aiming to train or fine-tune models. “Access to affordable computing power remains a critical bottleneck,” he explains.

Data acquisition is another challenge. To address this, Awarri launched LangEasy.ai, a platform collecting thousands of voice samples from participants in the government’s 3 Million Technical Talent initiative. Still, many researchers must gather their own specialized datasets, complicating project initiation and benchmarking.

Funding constraints further limit progress. Although the government has supported 20 peer-reviewed AI research papers over 18 months, this output is modest relative to Nigeria’s population exceeding 200 million. Without increased investment in grants, infrastructure, and compute access, N-ATLAS risks remaining a symbolic achievement rather than a transformative catalyst.

Moreover, while the open-source license permits free experimentation and prototyping, scaling products beyond 1,000 users requires formal licensing, which may hinder commercialization efforts without sustained financial backing.

Charting the Future: Collaboration, Inclusion, and Cultural Preservation

The ultimate value of N-ATLAS lies in fostering collective ownership of Nigeria’s AI trajectory. Joshua Firima believes that broad adoption and private sector collaboration will be key. “Integrating AI into everyday platforms like phone lines and WhatsApp, alongside incentives such as grants, will drive meaningful impact,” he says. He also stresses the importance of privacy safeguards, bias mitigation, and accountability to build public trust.

Cultural preservation is a core motivation behind N-ATLAS. Aizehi Itua, Awarri’s Vice President of Marketing and Communications, highlights the urgency: “With over 2,000 languages across Africa and less than 2% represented in AI, many risk extinction without digital inclusion.”

The success of N-ATLAS will ultimately depend on how effectively startups, researchers, and ecosystem stakeholders leverage it to create practical tools for classrooms, clinics, and farms nationwide.

We invite your feedback on this article and suggestions for future AI topics. Please share your thoughts through this feedback form.

Save the date! Moonshot by TechCabal returns to Lagos on October 15-16, featuring Africa’s top founders, tech innovators, and creatives. Secure your tickets at moonshot.techcabal.com.