New Report Highlights the Global South’s Role in Shaping Ethical AI
A new report from the data and AI company SAS and the Global Center on AI Governance challenges the image of the Global South as a region that is lagging behind in AI development. Instead, the report highlights crucial opportunities for these countries to take an active role in shaping the AI of the future, with investments in capacity building, digital infrastructure and inclusive governance.
Potential was already recognized in the United Arab Emirates when the announcement of $1 billion initiative to expand AI in Africa was made late last year. The number of programmers in the UAE as a leading country in AI adoption are surpassing 450,000 and many of them are from developing countries.
One of the key challenges highlighted in the report is the need for local language models that take socio-economic and cultural nuances into account. Middle Eastern countries have been working for several years to develop large language models (LLMs) in Arabic, resulting in solutions such as K2 Think launched in UAE, ALLAM 34B developed by HUMAIN in KSA, and Fanar in Qatar.
The question of how to avoid a growing global AI divide has taken up a lot of space on the international agenda in recent weeks, not least during the World Economic Forum in Davos. The report, Constraint to Capability: Flipping the Narrative on AI in the Global South, is authored by Dr Josefin Rosén, Principal Trustworthy AI Specialist at SAS Institute, together with Dr Rachel Adams and Selamawit Engida Abdella from the Global Center on AI Governance. It argues that structural constraints such as limited infrastructure, data gaps, and governance immaturity can be wielded into strategic advantages if addressed deliberately. For African markets, including South Africa, this carries practical implications for economic competitiveness, digital sovereignty, and long-term stability.
The global AI economy is accelerating. But much of today’s AI development remains shaped by Western-centric datasets, infrastructure, and governance models. That concentration risks embedding new forms of dependency and limiting emerging economies’ ability to shape AI systems that reflect their languages, social realities, and regulatory priorities.
“The global AI gap risks creating a new form of inequality. But if we direct investments to the right areas, such as AI skills, access to representative data, and inclusive governance models, the Global South can play an active and meaningful role in shaping the future of AI. It’s about shifting the perspective, from limitations to opportunities,” says Dr. Josefin Rosén, Principal Trustworthy AI Specialist at SAS Institute.
Skilled AI professionals are increasingly moving to wealthier countries or working remotely for companies based abroad. Higher salary prospects and underdeveloped local markets are driving this trend, which in turn hinders local innovation and long-term development.
In response, countries across Africa and South Asia are investing in training programs and technology hubs that offer attractive, locally relevant opportunities. Governments are also experimenting with innovation grants for startups, tax incentives, remote work support, and short-term fellowships to retain talent or encourage it to return. The simple truth is this: talent stays where real opportunities exist and where people have the ability to shape the future.
Countries that are not burdened by decades of legacy AI systems may have greater freedom to design governance frameworks from first principles. For policymakers and technology leaders, this means the opportunity to embed ethical safeguards, transparency requirements, and data sovereignty protections from the outset, rather than retrofitting them later. AI trajectory will be shaped by how quickly institutions translate policy ambition into scalable operational governance frameworks and skills programmes.
“AI sovereignty is not about isolation. It is about ensuring that AI systems reflect local contexts, languages, and values, and that countries retain meaningful control over the data and models that influence their economies. The choices made today around skills development, infrastructure, and governance will determine whether AI becomes a driver of resilience or a source of new dependencies.
One of the report’s central concerns is representativeness. AI systems trained predominantly on Western datasets risk producing biased or incomplete outcomes when applied elsewhere. In practical terms, this can affect everything from credit scoring and healthcare diagnostics to public service delivery. In multilingual and socioeconomically diverse societies unrepresentative models may entrench exclusion rather than reduce it.
The report, therefore, calls for investment in local data ecosystems, including the development of language models that reflect indigenous knowledge and local languages of the region. It also explores the responsible use of synthetic data as a mechanism to increase representation while protecting privacy and managing data scarcity. The report ultimately argues that inclusion in the AI economy is not a question of charity, but of global stability. AI systems that exclude large populations from participation or misrepresent their realities create systemic risks.
It entails investing in AI skills and research capacity, building representative datasets, embedding transparent and enforceable governance, and structuring partnerships that strengthen, rather than dilute, sovereignty. The window for shaping this trajectory is narrow. AI systems are already being deployed at scale across finance, public services, healthcare, and infrastructure. The decisions made now will determine whether African economies become rule-takers or rule-makers in the AI era.
As Rosén notes, “This is about shifting perspective from limitation to capability. With the right investments and policy decisions, countries across Africa can help define what responsible and inclusive AI looks like globally.” The question is no longer whether AI will reshape the economy. It is whether the country will help shape that transformation or simply import it.
