In a previous post from September 2023 (linked above), I discussed new findings from an Economist article describing how artificial intelligence could transform scientific research through tools such as literature-based discovery and so-called “robot scientists.” I also warned that these AI-driven advances could have unintended consequences. In particular, they may disproportionately benefit wealthy countries, widening existing global economic inequalities and undermining the UN Sustainable Development Goal of reducing inequality. Such growing disparities could, in turn, intensify migration pressures from poorer regions to richer ones.
More recently, a paper entitled “Who on Earth Is Using Generative AI?”, published in the Elsevier journal World Development, provides the first comprehensive global analysis of individual-level adoption of generative AI. The study finds that country-level generative AI usage intensity is strongly correlated with both the coverage and quality of digital infrastructure. It further shows a pronounced concentration of usage across income groups: in early 2024, middle-income economies accounted for roughly half of global ChatGPT traffic, while low-income countries together contributed less than one percent. https://www.sciencedirect.com/science/article/pii/S0305750X25003468
Generative AI risks locking in global inequality rather than breaking it. Unlike earlier digital revolutions that lowered barriers and allowed poorer countries to leapfrog stages of development, AI rewards those who already possess advanced infrastructure, skilled labor, and strong institutions, effectively raising the threshold for participation and pushing many low-income economies to the margins. As AI-driven gains concentrate in wealthy and upper-middle-income countries, global income and opportunity gaps will widen, intensifying migration pressures and economic fragility in already vulnerable regions. The result is not merely economic divergence, but rising financial instability and geopolitical tension that inevitably spill across borders. This is precisely what the world does not need as Donald Trump’s return to power amplifies geopolitical turbulence, undermines international coordination, and ushers in an era of policy volatility and institutional strain—at a moment when global cooperation is already under severe stress.
Preventing AI from entrenching global inequality demands forceful and forward-looking policy choices. Rich countries and major technology firms must stop treating AI as a private advantage and instead recognize it as a shared global capability. This means investing seriously in digital infrastructure and education in poorer regions, enabling meaningful technology transfer, and keeping core AI tools open, affordable, and adaptable to local needs. If governed with ambition and solidarity, AI can become a catalyst for shared prosperity rather than a force that condemns billions to permanent exclusion from progress.
Declaration of competing interests — In January 2019, I proposed a deliberately abrasive idea: that rich countries should admit one African migrant for every €250,000 looted from Africa by its kleptocrats over the past fifty years. Even a conservative estimate lays bare the scale of the crime, well over ten million Africans displaced by the workings of kleptocracy. These are not charity cases, but people whose admission wealthy nations are morally indebted to. By this standard, Switzerland alone would owe asylum to more than one million.
PS — In this context, it is important to recall a previous post highlighting how income inequality and poor living conditions can fuel terrorism and radicalization. It draws on a 2019 study by Krieger and Meierrieks, based on data from 113 countries, which identifies a clear link between inequality and terrorism. https://pacheco-torgal.blogspot.com/2019/09/2019-paper-by-german-researchersincome.html
