Inclusivity must be at the heart of AI development – Professor Marwala

Artificial intelligence (AI) is at the heart of this transformation, offering immense potential to innovate and address some of the world’s most pressing challenges, says Professor Tshilidzi Marwala, Rector of the United Nations University and Under-Secretary-General of the United Nations.

Marwala delivered the keynote address on day two of the 17th edition of the International Conference on Theory and Practice of Electronic Governance (ICEGOV), hosted by South Africa’s Department of Public Service and Administration, the Centre for Public Service Innovation and the Wits School of Governance (Wits University) in partnership with United Nations University.

In his address titled, Digital Cooperation – International Governance of Emerging Technologies & the Global Digital Compact, Marwala said as technological advancements are rapidly reshaping our economies, industries, and societies, difficult questions need to be asked. 

“How do we govern this rapidly evolving field responsibly and ethically? This question lies at the core of what I call the AI balancing problem. We must ask ourselves: how do we push the boundaries of innovation while ensuring that human values, rights, and dignity remain central? How do we foster progress while safeguarding against the risks AI presents,” said Marwala.

He emphasised that AI should be included in strategies for achieving the Sustainable Development Goals (SDGs) but must be done responsibly and should not exacerbate the inequalities these goals are seeking to eliminate.

“It is estimated that AI could contribute to 134 targets across 79% of all SDGs, accelerating progress on issues from poverty eradication to climate action. Data-driven development, for all its promise, brings new risks. 

“This year alone, over 2 billion people will have participated in democratic elections worldwide. Yet we have seen how technologies like deepfakes can undermine the integrity of these processes, posing serious threats to global democracy. The challenge, then, is clear: comprehensive governance frameworks are not just desirable; they are essential to protecting our systems and institutions,” he added.

Risks

While AI offers numerous benefits, the risks are far-reaching, especially biases.

“The biases ingrained in AI systems, particularly when trained on unrepresentative or skewed data. These biases can lead to discriminatory outcomes, often disproportionately affecting vulnerable populations,” said Marwala. 

The issue of language is something else that needs to be included in technology development and conversations.

“I worked with this firsthand when exploring the challenges faced by AI language recognition technologies in understanding low-resourced languages, such as Ju/’Hoansi San, spoken by an indigenous group here in southern Africa. Being spoken only by tens of thousands, this language is underrepresented in digital contexts, leading to AI systems that struggle to capture its nuances,” Marwala explained.

He said addressing bias in AI is both a technical and moral imperative and governments and institutions must take active steps to ensure that diverse datasets are used, and that countries with fewer resources have access to the tools they need to thrive in a digital world. "Inclusivity must be at the heart of AI development—this is not only an ethical issue but one of economic and social justice."

Global Digital Compact (GDC)

Marwala said at the recent Summit of the Future, UN Member States adopted the Pact for the Future, a global call for enhanced cooperation to address pressing global challenges and promote sustainable development through collaboration and innovation. 

“The Global Digital Compact is built on shared principles and a united commitment to bridge the digital divide while fostering an inclusive and safe digital environment. It calls for cooperation between governments, the private sector, civil society, and international organisations to create a digital future that benefits all.

“The Compact emphasizes the need to close the digital divide between the Global North and Global South, a divide that—if unaddressed—threatens to further entrench existing inequalities. It calls for gender-sensitive approaches and a renewed focus on ensuring that digital technologies are accessible to all, regardless of geography or socioeconomic status. These commitments come at a pivotal moment. Progress on the SDGs is stalling on multiple fronts, and the time to act is now. Emerging technologies, including AI, must be harnessed not as tools of exclusion but as catalysts for transformation, he said.

Examples of AI in action in Africa 

South Africa: AI-driven water management systems are already helping to mitigate the impacts of water scarcity. 

Tanzania: AI is revolutionising education through personalised tutoring systems. Rwanda: Their healthcare system has seen significant cost reductions through AI-enabled procurement systems. 

Across Africa and Asia, AI-powered credit scoring is enhancing access to financial resources for people with no formal credit history, helping to build more inclusive economies.

“When AI is developed within the Global South for the Global South, it becomes both contextually relevant and culturally appropriate. The key to enabling full participation in the digital economy lies in ensuring that developing countries have access to the technologies and resources they need to innovate and grow. Equally important is investing in youth education and equipping the next generation with the skills required to thrive in a rapidly changing world,” said Marwala.