Saheed Azeez, a student at the University of Lagos in southwest Nigeria, recently made headlines for developing a text-to-speech artificial intelligence model, YarnGPT, that reads text in distinctly Nigerian accents.

Azeez’s creation is one of the few indigenous AI models developed by Nigerians, primarily for the Nigerian market. In recent years, many young people in the country have invested time, energy, and modest personal resources in developing models that serve local communities that are often neglected and misrepresented in mainstream AI systems.

Azeez spent about 220,000 naira ($150) on the project. For an average Nigerian student, that is a significant amount of money. He self-sponsored the project partly because he wanted to avoid the pressure that comes with seeking to impress investors.

“I did try to get GPU sponsorship, but the process wasn’t straightforward. Eventually, I decided to build it myself without any external pressure, relying solely on my own funds. If I had investors, there would have been more pressure,” he told his school’s website.

Azeez’s experience highlights a broader challenge across the global south. For many people, the reality of building AI models that meet local needs is similar. Either there is a lack of institutional frameworks for training, grooming, and supporting talent capable of developing important AI tools, or there is general apathy toward creating the enabling environment required for AI development.

In countries like Ghana, despite a generally positive disposition toward AI, serious gaps remain in infrastructure and regulation. These two areas are critical for developing AI models and tools that address the needs of the local workforce.

Across much of the global south, AI ecosystems face significant hurdles that threaten inclusion and equitable access to its benefits. 

Popular and mainstream models and tools developed outside of the region often neglect significant cultural nuances.  Moreover, infrastructure gaps, particularly in rural and underserved regions, limit internet connectivity and access to digital tools essential for AI adoption.

In several countries, regulatory challenges persist. Laws lag behind technological advances, and fragmented approaches struggle to adequately address AI ethics, privacy, and inclusion. These gaps worsen exclusion risks for workers in regions already facing systemic marginalisation.

Data collection also poses a major challenge to AI development in Africa. Insufficient digital infrastructure, poor data quality, regulatory ambiguities, and a shortage of skilled professionals limit the availability of high-quality, localised datasets needed to build effective AI models.

Nigeria, for instance, has a somewhat chequered relationship with data. Fragmented government data silos and incomplete records from traditional collection methods hinder AI applications in sectors such as health and agriculture. Recent initiatives, including Google’s $2.25 million grant to modernise public data systems for AI readiness and Nigeria’s N-ATLAS AI model for local languages, aim to address these gaps by promoting data commons, interoperability, and open datasets.

I mentioned in an op-ed in October that Africa’s most populous nation is missing out on preparing its workforce for the future of work and failing to transition to modern technological systems. The same holds for many workforces in the global south. But this is a solvable problem if the right political will is matched with strategic and adequate investment in AI development.

Expanding reliable digital infrastructure nationwide would reduce geographic and economic barriers. Reshaping educational curricula and vocational training programmes to emphasise AI and digital skills would help build a more diverse and capable workforce. Strengthening inclusive policy design by embedding accessibility and fairness mandates within digital identity and AI regulations can protect marginalised groups. And fostering collaboration among government, the private sector, academia, and international partners can spur innovation tailored to local needs while safeguarding rights.

Such comprehensive interventions would democratise the benefits of AI and create meaningful economic opportunities for all workers.

Oluwaseyi Akintola is a Talent Management professional with over ten years of experience in organization development, performance management, learning, and change management. At the International Monetary Fund in Washington, D.C., she manages enterprise-wide learning initiatives, coaching, and leadership programs, driving strategic talent policies and advancing workforce development through innovation and continuous improvement.