Artificial Intelligence (AI) innovation in Nigeria and Africa may be driven by Small Language Models (SLMs), according to leading experts.
Since the launch of ChatGPT in November 2022, Large Language Models (LLMs) have demonstrated the vast potential of AI, inspiring platforms such as Google’s Gemini and Microsoft’s Co-Pilot. These advancements have expanded AI applications in generative text, speech, images, and video processing.
However, the extensive computational power and vast datasets required to develop and maintain LLMs make them inaccessible in many regions, particularly in emerging markets like Nigeria. Recognizing this challenge, Bosun Tijani, Nigeria’s Minister of Communications, Innovation, and Digital Economy, emphasized in 2024 the need for Nigeria to establish itself as a global player in AI development and regulation.
While LLMs like GPT-4 can exceed 175 billion parameters, SLMs typically range from tens of millions to under 30 billion parameters, according to a World Economic Forum (WEF) report. This makes SLMs more efficient, cost-effective, and practical for regions with limited infrastructure.
A white paper co-authored by Olubayo Adekanmbi, founder of Data Science Nigeria, and Ife Adebara, an AI expert, highlights how SLMs can drive AI innovation in Nigeria and across Africa. Their research suggests that SLMs offer a sustainable AI development pathway for emerging markets facing challenges such as limited digital infrastructure, offline access restrictions, and budget constraints.
According to Olivia Shone, Senior Director of Product Marketing at Microsoft, SLMs are specifically designed for targeted AI tasks, making them more efficient and resource-friendly. She explains that SLMs can generate responses with the same level of expertise as LLMs for domain-specific queries but at a lower cost and with reduced latency.
Experts argue that SLMs provide significant benefits for developing economies like Nigeria, including lower computational requirements thereby reducing infrastructure costs, cost-effectiveness, they can be easily tailored to meet local and industry-specific needs and can also function in areas with limited or no internet access.
According to Adekanmbi and Adebara, who co-founded EqualyzAI, SLMs can democratize AI access by enabling governments, small businesses, and individuals to integrate generative AI into their daily operations.
In mobile-first economies like Nigeria, where internet connectivity remains inconsistent, SLMs provide an alternative that works on low-power devices and offline environments. This ensures that rural populations and areas can participate in AI-driven advancements.
AI experts such as Libing Wang, Chief of Section for Education at UNESCO Bangkok, and Tianchong Wang, Lecturer at Swinburne University of Technology, Australia, stress that SLMs provide an opportunity for Global South nations to overcome digital infrastructure constraints. By developing AI applications tailored to local needs, SLMs can accelerate technological transformation across multiple sectors.
“SLMs have the potential to revolutionize accessibility, efficiency, and affordability in AI adoption,” they argue. Despite their advantages, SLMs are not without limitations. The World Economic Forum points out that SLMs may struggle with complex linguistic tasks, have lower accuracy in certain contexts, and exhibit a narrower performance scope than their larger counterparts.
However, as AI continues to evolve, experts believe that SLMs will play a crucial role in enabling developing nations to leverage AI technology effectively while remaining economically sustainable.
