According to the "Technology and Innovation Report 2025," officially launched this April 7 by the United Nations Conference on Trade and Development (UNCTAD), the AI market could reach $4.8 trillion by 2033, an amount equivalent to Germany's gross domestic product. The UN agency also anticipates that 40% of jobs worldwide could be affected by AI-induced automation, a projection that echoes estimates from the International Monetary Fund. Both also warn of a risk of widening the digital divide between advanced economies and developing countries.
UNCTAD, a UN body that celebrated its 60th anniversary last year, was created to help developing countries better integrate into the global economy. It aims to promote fairer and more inclusive trade by relying on policies that support economic growth and sustainable development.
The 2025 edition of its series of reports on technology and innovation, initiated in 2010, titled "An Inclusive Artificial Intelligence for Development," seeks to guide policymakers through the complex landscape of AI and help them implement public policies that ensure more equitable technological growth. It identifies three essential levers for this: infrastructure, data, and skills.
A Dynamic of Technological Concentration
Among its conclusions, the report highlights the concentration of AI investments: nearly 40% of global research and development (R&D) expenditures are currently in the hands of around a hundred companies, mainly based in the United States and China. This concentration strengthens a dynamic of technological monopoly, where countries with advanced digital ecosystems see their competitiveness increase, while emerging economies struggle to develop their own infrastructures and skills.
In this context, major industry players such as Nvidia, Microsoft, or Apple display market capitalizations comparable to, or even exceeding, the GDP of the entire African continent. This imbalance goes far beyond financial scope: it reflects structural gaps in terms of infrastructure, talent, and digital sovereignty, which durably influence economic power dynamics.

Automation: A Differentiated Impact Across Regions

Office and administrative jobs will be the most affected by AI, mainly impacting advanced economies. In low-income countries, where low-cost labor has been a competitive advantage, this proportion is lower, but the indirect effect could be more destabilizing.
However, UNCTAD emphasizes the need to nuance this finding: AI, if accompanied by active public policies in training, local innovation, and equitable access to data, can also foster the emergence of new sectors and strengthen productive capacities.

Towards More Inclusive AI Governance

The report also highlights that about 118 countries, mainly from the Global South, remain absent from global discussions on AI governance, thereby reinforcing the dominance of major digital powers.
Faced with these risks of technological exclusion, the report advocates for more inclusive global governance. It notably proposes the establishment of a public disclosure mechanism for AI uses, the development of shared infrastructures, and support for open-source models. It also calls for increased international cooperation to limit technological disparities.
Concluding:
"AI can be a catalyst for progress, innovation, and shared prosperity, but only if countries actively shape its trajectory."

Translated from Alors que 40 % des emplois pourraient être impactés par l'IA, la CNUCED appelle à une coopération internationale

To better understand

What are the main technical challenges related to the infrastructure needed to support inclusive AI?

Technical challenges include the availability and quality of storage and computing infrastructures, access to data, and the ability to process massive amounts of information in real-time. Establishing robust networks for connectivity and cybersecurity is also crucial.

How could global AI governance evolve to be more inclusive?

To be more inclusive, global AI governance could adopt transparency mechanisms on AI usage, encourage more intense international collaborations, and support open-source models. It would also involve the active participation of developing countries in global discussions.