While the Fifa World Cup captivates billions with its closely fought matches and intense rivalries, the results of a far more consequential contest were quietly published this month. Its top performer is not a football powerhouse like Brazil, Germany or England. It is the UAE – a young nation of a little more than 10 million people that did not exist until 1971.
Microsoft’s AI Economy Institute released its quarterly Global AI Diffusion Report – the first of this year – tracking how widely generative AI has penetrated societies and economies worldwide. The UAE ranked first globally, with 70.1 per cent of its working-age population actively using generative AI – up from 64 per cent just a quarter earlier, which was also the first globally, and from 59.4 per cent a year ago.
Not only did the UAE top the table, the 6.1 percentage-point quarterly gain was among the largest of any leading nation. For context, second-placed Singapore sits at 63.4 per cent, Norway, Ireland and France – third, fourth and fifth – cluster about 48 per cent, while the US, widely assumed to be an AI superpower, languishes at 31.3 per cent and in 21st place, having crept up only three spots in a quarter. The global average is 17.8 per cent.
Reaching the top of a league table is one thing. Widening the gap once you are there is another. This did not happen by accident, nor by infrastructure alone.
Everyone knows that the UAE was ahead of the curve and appointed the world’s first Minister of State for Artificial Intelligence in 2017 – before most governments had started grappling with the question of how AI could affect all aspects of life. Soon after that, a comprehensive national strategy followed, and we started seeing eye-watering investments in digital infrastructure, cloud capacity, data ecosystems and research institutions sustained over nearly a decade. The difference between the UAE and most of its peers is not resources – it is sequencing, speed and a clear vision of the future.
In the new age of competitive advantage of nations, the UAE is now in prime position as history offers a useful lesson. Technological revolutions rarely reward those who invent a technology first. They reward those who diffuse it fastest. Britain invented the railway; America built a continent with it. Europe shaped the internet’s architecture; Silicon Valley colonised its economics. The country that leads in AI adoption – not merely AI research – will compound advantages across productivity, health care, education, government efficiency and economic growth. The UAE has grasped this with unusual clarity.
Elsewhere, the report reveals a story with two distinct subplots.
The first is Asia’s emergence as a new growth engine. Twelve of the 15 fastest-growing economies since last June are Asian. South Korea leads the surge with a 43 per cent increase in AI users over that period, now at 37.1 per cent adoption – a jump that lifted it from 18th to 16th in the global rankings. Thailand is up 36 per cent, Japan 34 per cent.
The driver is not demographics or wealth alone, but something more specific: rapidly improving multilingual AI capability. As models close the gap between English and local-language performance – Japanese professional exam accuracy has risen from about 51 per cent on early GPT models to more than 90 per cent on recent systems – adoption follows almost mechanically. Language was the lock; better models are the key.
The second subplot is less encouraging. The gap between the Global North and Global South is not merely persisting – it is accelerating. In the first quarter of this year, the Global North reached 27.5 per cent AI adoption; the Global South, 15.4 per cent. A year ago, the gap was 9.8 percentage points. It is now 12.1.
The culprits are familiar: unreliable electricity, limited internet penetration, weak digital skills. The Global South has 88.9 per cent electricity access compared to 98.1 per cent in the North; internet access stands at 65.7 per cent versus 90.1 per cent. Until those foundations are addressed, AI’s gains will remain as unevenly distributed as the infrastructure that carries them.
The report also illuminates something happening at the productive frontier of the AI economy: a step-change in how software gets built. Git pushes – the measure of code uploaded to GitHub – rose 78 per cent year over year globally in the first quarter of this year, reaching 380 million. New repositories grew 45 per cent. More striking still, GitHub pull requests associated with AI coding agents grew 28-fold in just 10 months.
The alarm about AI destroying software jobs has, for now, proved premature: American software developer employment hit a record 2.2 million last year, up 8.5 per cent year on year, with early data from this year showing a further 4 per cent gain. When AI cuts the cost of building software, it appears organisations respond by building vastly more of it, and hiring more people to do that.
There is a caveat worth noting. Diffusion statistics measure usage, not sophistication. A population using generative AI to draft emails and summarise documents is meaningfully different from one producing AI-enabled medical breakthroughs or sovereign foundation models. The UAE remains a buyer and deployer of frontier technology. Turning adoption into indigenous innovation – closing the implementation-invention gap – is the critical challenge ahead.
Most economic value over the next decade will accrue not from building models but from deploying them at scale across industries and governments. On that measure, Abu Dhabi is lapping the field.
For policymakers elsewhere, the data should prompt discomfort rather than dismissal. Countries that continue to debate AI while others embed it are not being prudent, they are falling behind – and the gap, as this report makes plain, will get harder to close.
