Keir Starmer is keen to make the UK a global leader in AI. His government’s AI Opportunities Action Plan, unveiled in January, outlines a strategy to achieve that goal. The plan is to promote investment in digital infrastructure, establish digital sovereignty for the UK and ensure regulators work efficiently and avoid blocking innovation. That would be a fine start. But the competition for global AI leadership is intense. Governments around the world are vying for supremacy in cyberspace and some are far more likely than others to come out on top.
The Global AI Index, an annual report produced by Tortoise, a publisher, provides an overview of the global AI operating environment and seeks to explain which countries are best positioned to take advantage of the global AI boom – and why.
The research is complex. It measures each nation’s ‘AI capacity’, which is essentially their ability to contribute to and capitalise on the global AI boom. To determine AI-capacity scores, the researchers examined data covering 122 different indicators of AI readiness. They divided these into three scoring categories – implementation, innovation and investment – and, within those, seven sub-categories – talent, infrastructure and operating environment; research and development; and government strategy and commercial activity, respectively.
Since the research launched in 2019, the US and China have consistently been ranked first and second for AI capacity – and the third-place nation is never close. For years, it was the UK in third place, until Singapore forced it into fourth in 2023. Even between the perennial top two, however, the gap in capacity is significant.
The US ranks first or second in every sub-category and holds an overall capacity score of 100 – comfortably ahead of China, at 54. China’s score is dragged down by the country’s restrictive operating environment, relative lack of AI talent and government strategy. Italy ranks first for operating environment, which looks at factors such as legislative activity and public trust, and Saudi Arabia has the best government strategy, according to the research.
The ranking becomes more competitive starting at third place. Singapore (32), the UK (30) and France (28) rank third, fourth and fifth for AI capacity. The UK scores relatively highly across most sub-categories, but ranks just 16th for development, which is based on patent data and strength of the open-source community, as well as being 17th for infrastructure.
The chart above plots the sub-category scores of the top four countries by AI capacity. The US sets the standard for nearly every measure, with China coming in second in four of the seven sub-categories. Third-place Singapore has a considerably weaker operating environment than the UK, in fourth place overall, but ranks above the UK in infrastructure, research, development and commercial activity.
So China and the US are the clear leaders when it comes to AI capacity – but that’s perhaps not surprising. Many of the 122 indicators that underpin the research are measures of absolute output: the total government investment, the number of patent applications, the number of startups and company acquisitions and so on. Larger, wealthier, more populous countries have an advantage in many of these rankings.
How then do Singapore or Israel break the top 10 for overall capacity? The research measures not only scale – the absolute measure of AI capacity – but also intensity, which is the measure of a country’s AI capacity relative to the size of its population or economy.
Singapore, a nation of about 6 million people, ranks first for AI intensity, with a score of 100. Israel and Switzerland score 74 and 63, respectively. Despite its size, the US has a relatively high intensity score (73); however, China, with its 1.4 billion citizens, scores just 33 for AI intensity.
Overall capacity scores, however, are weighted toward scale (absolute) measurements, not intensity (relative) measurements. According to the report, AI capacity is about 75% scale and 25% intensity. The competition for AI supremacy, therefore, will likely remain a two-horse race.
Keir Starmer is keen to make the UK a global leader in AI. His government’s AI Opportunities Action Plan, unveiled in January, outlines a strategy to achieve that goal. The plan is to promote investment in digital infrastructure, establish digital sovereignty for the UK and ensure regulators work efficiently and avoid blocking innovation. That would be a fine start. But the competition for global AI leadership is intense. Governments around the world are vying for supremacy in cyberspace and some are far more likely than others to come out on top.
The Global AI Index, an annual report produced by Tortoise, a publisher, provides an overview of the global AI operating environment and seeks to explain which countries are best positioned to take advantage of the global AI boom – and why.
The research is complex. It measures each nation’s ‘AI capacity’, which is essentially their ability to contribute to and capitalise on the global AI boom. To determine AI-capacity scores, the researchers examined data covering 122 different indicators of AI readiness. They divided these into three scoring categories – implementation, innovation and investment – and, within those, seven sub-categories – talent, infrastructure and operating environment; research and development; and government strategy and commercial activity, respectively.
Since the research launched in 2019, the US and China have consistently been ranked first and second for AI capacity – and the third-place nation is never close. For years, it was the UK in third place, until Singapore forced it into fourth in 2023. Even between the perennial top two, however, the gap in capacity is significant.
The US ranks first or second in every sub-category and holds an overall capacity score of 100 – comfortably ahead of China, at 54. China’s score is dragged down by the country’s restrictive operating environment, relative lack of AI talent and government strategy. Italy ranks first for operating environment, which looks at factors such as legislative activity and public trust, and Saudi Arabia has the best government strategy, according to the research.
The ranking becomes more competitive starting at third place. Singapore (32), the UK (30) and France (28) rank third, fourth and fifth for AI capacity. The UK scores relatively highly across most sub-categories, but ranks just 16th for development, which is based on patent data and strength of the open-source community, as well as being 17th for infrastructure.
The chart above plots the sub-category scores of the top four countries by AI capacity. The US sets the standard for nearly every measure, with China coming in second in four of the seven sub-categories. Third-place Singapore has a considerably weaker operating environment than the UK, in fourth place overall, but ranks above the UK in infrastructure, research, development and commercial activity.
So China and the US are the clear leaders when it comes to AI capacity – but that’s perhaps not surprising. Many of the 122 indicators that underpin the research are measures of absolute output: the total government investment, the number of patent applications, the number of startups and company acquisitions and so on. Larger, wealthier, more populous countries have an advantage in many of these rankings.
How then do Singapore or Israel break the top 10 for overall capacity? The research measures not only scale – the absolute measure of AI capacity – but also intensity, which is the measure of a country’s AI capacity relative to the size of its population or economy.
Singapore, a nation of about 6 million people, ranks first for AI intensity, with a score of 100. Israel and Switzerland score 74 and 63, respectively. Despite its size, the US has a relatively high intensity score (73); however, China, with its 1.4 billion citizens, scores just 33 for AI intensity.
Overall capacity scores, however, are weighted toward scale (absolute) measurements, not intensity (relative) measurements. According to the report, AI capacity is about 75% scale and 25% intensity. The competition for AI supremacy, therefore, will likely remain a two-horse race.
