A tech entrepreneur leading the African-centred data initiative, DataLens Africa, has emphasised the urgent need for well-organised and culturally relevant datasets across Africa, warning that the continent may fail to maximise the opportunities presented by artificial intelligence if its data collection system remains disorganised and fragmented.
In an interview with AIbase Nigeria, Olaoye Anthony Somide, Founder and CEO of CipherSense AI, explained that Africa still struggles with data accessibility and fragmentation despite producing huge volumes of information every day.
Somide stated that the continent’s main issue is not a shortage of data, but rather poor coordination, lack of organisation, and limited accessibility, all of which reduce the effectiveness of data for innovation, research, policymaking, and AI development.
“Africa has a lot of data, but most of it is scattered and poorly organised. It is like a flood without structure. The information exists, but accessing and using it efficiently remains a challenge,” he said.
He explained that DataLens Africa, operating under CipherSense AI Africa, is committed to creating and organising Africa-focused datasets across industries such as healthcare, education, agriculture, finance, language, and culture. According to him, the objective is to ensure AI systems designed for Africans accurately reflect local realities and everyday experiences.
Somide added that the initiative is supported by a growing network of contributors across several African countries, with participants trained through webinars, onboarding programs, and community engagement activities.
He noted that the organisation maintains active communication with contributors through regular training sessions that expose them to real project workflows, operational procedures, and data quality standards.
In addition, the company conducts onboarding sessions whenever new AI or data projects are obtained, ensuring contributors fully understand project expectations, workflows, and deliverables.
To maintain long-term collaboration, Somide said the organisation uses platforms like Discord and LinkedIn, where contributors interact, ask questions, and stay updated on projects and opportunities.
He also expressed concerns about Africa’s dependence on foreign-trained AI systems, noting that many global AI models struggle to adequately understand African social and cultural realities.
According to him, AI systems trained primarily on Western datasets often produce responses that do not align with African lifestyles and behavioural patterns.
“For example, if you ask some global AI systems for birthday gift ideas, they may suggest flowers or other Western-centric options. But an AI system trained on the African context may suggest sending money, making a phone call, or other culturally relevant gestures,” he said.
Somide further explained that poor AI performance in some regions is often due to deploying unsuitable AI systems rather than to weaknesses in the technology itself.
“Sometimes the problem is not that AI is bad. The problem is deploying the wrong AI for the local environment,” he stated.
Related:
- 8 Top Websites That Pay People to Train AI Models in 2026
- Anthropic Unveils Mythos AI Model in New Cybersecurity Push
- Fragmented Data Rules Threaten Africa’s AI Ambitions: Microsoft
The CipherSense AI CEO also highlighted the importance of quality data in developing dependable AI systems, stressing that accurate annotation and strict quality control are essential for strong model performance.
“AI systems rely heavily on properly labelled data. Even small errors in data handling can significantly affect model outcomes,” he explained.
He revealed that CipherSense AI has assembled dedicated teams comprising quality-assurance specialists, AI evaluators, and African-language experts to ensure that datasets accurately reflect African communication patterns and cultural diversity.
According to him, using local expertise in data collection and annotation improves the reliability and relevance of datasets compared to more generalised global methods.
Speaking on AI safety concerns, Somide warned that many global AI systems still contain biases inherited from their training data, including stereotypes related to geography and identity.
“One of the major issues with some global AI models is bias. An AI system could mistakenly associate certain regions with fraud or crime due to biased data. These are issues we actively work to identify and reduce,” he said.
He also discussed the potential of localised AI systems to improve healthcare delivery in Africa through diagnostic tools trained on region-specific disease patterns.
Somide mentioned diseases such as malaria and Ebola, which are widespread in Africa but often underrepresented in global healthcare datasets.
“If AI systems are trained using African healthcare data, they can support faster and more accurate diagnosis for diseases common in our environment,” he said.
This discussion comes as calls continue to grow for Africa to leverage locally generated, well-structured data to address its unique challenges. Industry experts increasingly believe that culturally relevant datasets will play a critical role in building fair, efficient, and context-aware AI systems capable of solving real problems across the continent.
Senior Reporter/Editor
Bio: Ugochukwu is a freelance journalist and Editor at AIbase.ng, with a strong professional focus on investigative reporting. He holds a degree in Mass Communication and brings extensive experience in news gathering, reporting, and editorial writing. With over a decade of active engagement across diverse news outlets, he contributes in-depth analytical, practical, and expository articles exploring artificial intelligence and its real-world impact. His seasoned newsroom experience and well-established information networks provide AIbase.ng with credible, timely, and high-quality coverage of emerging AI developments.