While the world’s focus is on the killing fields in Iran and Lebanon, two major, peaceful geopolitical developments have occurred in New Delhi and Beijing.
On February 19-21 the Indian government hosted the fifth major summit on AI. Previous ones were hosted by the UK government in November 2023 focusing on AI safety, in Seoul in May 2024, where innovation was added to the theme of AI safety, and in Paris in February 2025, which centred AI action.
In April 2025, Rwanda hosted the Global AI Summit on Africa, which twinned the themes of AI and Africa’s demographic dividend. Last November, at the G20 in South Africa, the AI task force affirmed that jurisdictions must develop and regulate AI in terms of their own legal frameworks, referred to as “digital sovereignty”, while committed to risk-based, human-centric and development-oriented AI policy approaches.
This served as a precursor to India’s summit theme, AI Impact, with the summit declaration emphasising that AI’s benefits should be broadly shared by all of humanity.
Barely two weeks later, on March 5, the Chinese unveiled their 15th five-year plan. As The Ascent Begins: The World Beyond Empire - Sovereignty in the Age of Collapse author Shanaka Anslem Perera has pointed out, the 141-page plan mentions AI more than 50 times, with AI expected to affect 90% of the economy by 2030, and China aiming to reach a value of $1.38-trillion in AI-related industries covering virtually every sector – robotics, space infrastructure, quantum computing and rare earth processing.
We need to locate this in the context of the geopolitics of AI: the US controls the upper end of the AI value chain in advanced computing, while China controls much of the critical minerals supply chain that underpins electronics manufacturing.
Perera argues that China’s rare earth dimension is where its plans intersect with the US’s war plans. With China controlling 90% of global rare earth processing, he contends that every F35 joint strike fighter, every Patriot missile battery, every THAAD interceptor, every guided munition, depends on materials China processes.
What are the implications of these developments for South Africa? Having just concluded the presidency of the G20 with AI equity as a central theme, we need to position ourselves as a credible interlocutor between the India-led Global South AI agenda and African interests.
We need to find our place among the “tech middle powers” alongside countries such as India, Indonesia, Malaysia, Singapore, Australia and Brazil. We should forge partnerships to promote AI development and governance, generate new cooperation streams for AI standards and frameworks, and co-develop AI capabilities.
Any developments in technology must guard against a built-in Western-centred bias, such as in facial or language recognition systems. Among the concrete examples we should look at is BharatGen, India’s first government-supported multimodal AI model, designed to operate across 22 Indian languages with text, speech and image capabilities.
Aadhaar, India’s digital ID system with over 1-billion users, has been studied by South Africa’s officials involved in MyMzansi, which aims to achieve digital transformation in the public sector. Both systems need to heed concerns expressed about reinforcing existing inequalities and access to the services promised via AI.
The Indian summit marked the launch of the Evidence for AI in Health initiative, a major global effort to generate rigorous evidence on responsible AI use in health systems across low- and middle-income countries in sub-Saharan Africa, South Asia and Southeast Asia. Supported by the Wellcome Trust, the Gates Foundation and the Novo Nordisk Foundation, it creates a specific entry point for South African health researchers and institutions, with April 1 being the deadline for responding to the request for proposals.
The next step globally and locally is translating the various norms emerging from these summits into domestic legislation. South Africa released a national AI policy framework in October 2024, but a comprehensive AI-specific legal framework is still needed to address accountability and ethical concerns around AI deployment.
We also need to stop being the nice guys spouting all the correct rhetoric and become a node for actual AI capability on the continent.
• Abba Omar is director of operations at the Mapungubwe Institute for Strategic Reflection.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.