AI: Africa and Latin America combined account for 3.3% of the world's data centers, warns the ITU

Africa and Latin America combined account for only 3.3% of global data center capacity, stated Tomas Lamanauskas, Deputy Secretary-General of the International Telecommunication Union (ITU), in an interview published by Al Jazeera on October 25, 2026, on the sidelines of the UNCTAD World Investment Forum in Doha. The ITU is the United Nations agency responsible for information and communication technologies.

For this official, the lag is no longer just a digital divide; it is becoming an "opportunity gap" because AI impacts health, education, public services, industry, agriculture, and trade. Globally, 2.2 billion people still lack access to the internet, and closing the digital infrastructure gap would require approximately $2.8 trillion in investment by 2030, according to the figures he provided.

Our languages absent from machines

One figure should capture the attention of African capitals: only 7% of African languages are represented in AI translation tools, again according to the ITU. Technology that does not speak Wolof, Amharic, Hausa, or Kiswahili at the same level as English does not primarily serve those who speak them. Usage gaps follow: approximately 29% of workers in developed countries use generative AI, compared to 16% in developing countries.

Jobs to lose, others to gain

AI could eliminate 90 million jobs and create 170 million, and nearly 40% of professional skills could evolve by 2030, the ITU indicates. A report by the United Nations Development Programme published in December 2025 warns that the technology could usher in a new era of inequality between nations and erase decades of progress: countries with unreliable electricity, limited connectivity, or rare technical skills risk missing out on the gains.

The international report on AI safety from February 2026, cited by Al Jazeera, estimates the number of users of these tools at around one billion and considers 60% of jobs in advanced economies and 40% of those in emerging economies to be exposed, while acknowledging that data is lacking. Fazl Barez, a researcher at the University of Oxford, distinguishes between the automation of certain tasks and the replacement of entire jobs: repetitive tasks are the most threatened, but the scale remains uncertain.

There is still room for maneuver. One hundred and four countries now have a national AI strategy, compared to four a decade ago. Lamanauskas believes it is impossible to stop this technology and calls for investment in infrastructure, education, and regulation so that developing countries participate in its design and not just its consumption. For the continent, the challenge is clear: host its own data, train its engineers, and integrate its languages into the models.