Nigeria has unveiled a continent-wide artificial intelligence (AI) initiative aimed at ensuring the next generation of AI systems can understand, speak and serve Africa’s thousands of indigenous languages, marking a major step towards building sovereign AI capabilities across the continent.
The initiative, known as the ATLAS Network, was launched during the opening ceremony of the 7th Ordinary Session of the African Telecommunications Union (ATU) Conference of Plenipotentiaries (CPL-26) in Abuja, where African leaders identified artificial intelligence as a strategic pillar for the continent’s digital transformation rather than simply another emerging technology.
Building on Nigeria’s pioneering N-ATLAS initiative, the new continental network seeks to ensure that Africa’s languages, cultures and knowledge systems are embedded in the next generation of AI technologies, enabling AI systems to better understand indigenous languages, regional accents, cultural contexts and local realities.
The initiative comes as governments worldwide race to develop sovereign AI capabilities amid growing concerns over language bias, data ownership and the dominance of AI models trained largely on Western datasets.
According to Nigeria’s Federal Ministry of Communications, Innovation and Digital Economy, the ATLAS Network will bring together governments, universities, researchers, innovators and ecosystem partners across Africa to accelerate the development of locally relevant Large Language Models (LLMs) capable of serving African societies.
Building on Nigeria’s pioneering N-ATLAS initiative, the new continental network seeks to ensure that Africa’s languages, cultures and knowledge systems are embedded in the next generation of AI technologies, enabling AI systems to better understand indigenous languages, regional accents, cultural contexts and local realities. The initiative comes as governments worldwide race to develop sovereign AI capabilities amid growing concerns over language bias, data ownership and the dominance of AI models trained largely on Western datasets.
Closing AI’s African language gap
For decades, most of the large language models powering today’s AI assistants have been trained predominantly on English and other major global languages using datasets originating largely from North America and Europe.
While these models perform strongly in widely spoken international languages, they often struggle with African accents, indigenous languages, local place names, cultural expressions and context-specific knowledge.
This imbalance has significant implications as AI becomes increasingly integrated into education, healthcare, agriculture, financial services, public administration and digital government platforms across Africa.
Experts warn that unless African languages are adequately represented in AI training datasets, more than 2,000 indigenous languages across the continent risk remaining digitally invisible, limiting access to AI-powered services for millions of people.
Translation systems frequently produce inaccurate outputs for many African languages, voice assistants often fail to recognise local pronunciations, while AI-powered educational and healthcare applications can struggle to understand cultural nuances unique to African communities.
Although researchers across Africa have begun addressing these challenges through language digitisation projects and indigenous AI research, Nigerian officials say the ATLAS Network represents one of the continent’s boldest collaborative efforts yet to systematically bridge the AI language divide.
From Nigeria’s N-ATLAS to Africa’s ATLAS Network
The continental initiative builds directly on Nigeria’s N-ATLAS (Nigeria Atlas for Languages and AI at Scale) programme introduced in 2025 as Africa’s first government-backed open-source multilingual large language model initiative.
N-ATLAS was designed to support AI applications in Nigerian languages, beginning with major indigenous languages alongside Nigerian-accented English, while providing an open platform that researchers and developers could continuously improve.
The newly launched ATLAS Network significantly expands that vision beyond Nigeria by creating a collaborative continental ecosystem where African governments, universities, research institutions and technology communities jointly develop language models that reflect Africa’s linguistic diversity, cultural heritage and indigenous knowledge systems.
Rather than each country independently developing AI language resources, the initiative promotes cross-border collaboration that allows participating nations to pool expertise, computing resources, datasets and research capabilities.
Officials believe this collaborative approach could accelerate the development of AI systems capable of serving millions of Africans in their native languages while reducing dependence on foreign AI platforms.
Tijani: Africa must shape AI, not merely consume it
Announcing the initiative, Dr Bosun Tijani, Minister of Communications, Innovation and Digital Economy, described the conference as a defining moment for Africa’s digital future and called for stronger continental cooperation as artificial intelligence reshapes economies around the world.
“The Conference of Plenipotentiaries is where Africa reflects on its progress and charts a common course for the future. At a time when digital technologies are reshaping economies and societies at an unprecedented pace, our responsibility to work together has never been greater,” Tijani said.
“Today, we gather to strengthen the partnerships and shared resolve that will ensure Africa continues to play a leading role in shaping the future of telecommunications and the digital economy.”
The initiative comes as governments worldwide race to develop sovereign AI capabilities amid growing concerns over language bias, data ownership and the dominance of AI models trained largely on Western datasets. Image credit: FMCIDE.
Beyond developing African AI models, the minister linked the initiative to the continent’s broader participation in international digital governance and technology policymaking. According to him, Africa must become an active contributor to global technology standards rather than simply adopting frameworks developed elsewhere.
Beyond developing African AI models, the minister linked the initiative to the continent’s broader participation in international digital governance and technology policymaking.
According to him, Africa must become an active contributor to global technology standards rather than simply adopting frameworks developed elsewhere.
“Africa must arrive prepared, united and credible. We must continue to invest in technical preparation, strengthen regional cooperation and speak with one voice where our interests align,” the minister added.
“Where global digital rules are being shaped, Africa must be ready to participate and help shape outcomes that affect our people, our economies and our future.”
His remarks reflect growing recognition among African policymakers that the continent risks becoming underrepresented in the rapidly evolving AI ecosystem unless it invests in indigenous datasets, local computing infrastructure and multilingual AI research.
Why language matters in AI
Language has emerged as one of the most important frontiers in artificial intelligence.
Large Language Models learn from enormous volumes of digital text. Where languages lack sufficient digital content, AI systems cannot effectively understand or generate meaningful responses in those languages.
For Africa, whose linguistic diversity exceeds 2,000 languages, the challenge is particularly significant.
Without deliberate investment in African language resources, AI systems could inadvertently deepen existing digital inequalities by excluding millions of citizens from accessing intelligent digital services in languages they understand best.
Conversely, AI models trained on African languages could improve translation services, voice assistants, education technologies, healthcare applications, agricultural advisory systems, financial inclusion platforms and digital public services tailored to local communities.
By embedding African languages, cultures and indigenous knowledge into foundational AI models, governments hope to ensure that future AI technologies better reflect the realities of African societies rather than relying solely on perspectives embedded in foreign datasets.
Strengthening Africa’s digital sovereignty
Beyond language preservation, the ATLAS Network represents a broader effort to strengthen Africa’s digital sovereignty as governments increasingly view AI infrastructure as strategic national infrastructure.
The initiative aims to encourage African countries to jointly develop sovereign AI capabilities while retaining greater control over the datasets, research and technological resources that underpin next-generation AI systems.
Officials believe closer collaboration between governments, universities, researchers and innovators will enable Africa to build AI technologies that are more inclusive, culturally relevant and better aligned with the continent’s development priorities.
As artificial intelligence becomes a defining technology of the global digital economy, Nigeria’s launch of the ATLAS Network signals an ambition not only to expand AI adoption across Africa, but also to ensure that future AI systems can speak African languages, understand African cultures and solve uniquely African challenges.
Home
