A IMGlobalWealth.news report
A warning that artificial intelligence could widen rather than narrow economic inequality is gaining weight as new evidence shows the technology’s early benefits flowing disproportionately to richer countries, higher-paid workers and asset owners.
The Washington Post reported on Thursday that Nobel laureate Daron Acemoglu and executives at Bridgewater Associates are among those warning that the AI boom risks reinforcing existing economic divides. The concern is not that AI is making everyone poorer. Rather, it is that the gains are arriving much faster at the top.
“It is that AI is generating substantial new wealth, but access to the tools, skills and capital needed to capture it remains highly unequal”
That distinction is supported by recent international data. An IMF working paper published in July, drawing on AI usage across more than 100 countries, estimated the annualised labour-cost value of time saved by AI at about $2.7trn across 86 economies, equivalent to 3.4% of their combined GDP. Yet 96% of that measured value accrued in high-income countries. Relative to GDP, the gain was estimated at 4.2% in rich economies, around 0.6% in middle-income countries and just 0.1% in low-income ones.
The same research found AI gains tilted towards higher-paid occupations in almost every country examined. That does not mean wages will diverge by the same amount, since productivity gains can accrue to workers, companies or consumers. But it does show where the technology is currently creating value.
Asset ownership creates a second divide. Federal Reserve data show that US households in the wealthiest 10% held about $48trn in corporate equities and mutual funds in the first quarter of 2026, compared with roughly $7trn for the remaining 90%. As AI-related companies have become major drivers of equity-market wealth, the resulting wealth effect is inherently uneven.
Labour-market evidence is also becoming harder to dismiss. A Stanford Digital Economy Lab study updated this month found no economy-wide AI jobs collapse, but employment among 22-to-25-year-olds in AI-exposed occupations stood 19% below where it would have been had it kept pace with less-exposed work. The divergence largely reflected weaker hiring rather than mass redundancies.
There is, however, a counterweight. AI can make less-experienced workers more productive, lower the cost of expertise and help smaller firms compete. IMF data also show the concentration of AI gains declining in many countries as adoption spreads.
For now, the emerging story is not simply that AI makes the rich richer and the poor poorer. It is that AI is generating substantial new wealth, but access to the tools, skills and capital needed to capture it remains highly unequal. Whether that gap narrows may become one of the defining economic questions of the AI age.
