The World Bank on Tuesday warned that developing countries have a narrow window to seize on the economic gains that artificial intelligence can deliver.
In its latest World Development Report, the World Bank said countries that fail to act could increase AI inequality, concentrate market power and weaken trust in public institutions.
The World Bank drew comparisons of AI's emergence to that of other game-changing historical advancements such as steam power and the internet, the gains of which transformed the entire economic spectrum.
“This is a golden opportunity, and if developing countries don't put the policies in place that can help them build the complements or the foundations that are needed to benefit from AI, then there is a risk that they will fall behind,” said Gaurav Nayyar, co-author of the report.
While developing AI infrastructure and related tools can cost billions of dollars, developing countries can use adapt low-cost tools to local conditions, to improve services such as medical care, agriculture and education.
“Be optimistic, be strategic, but also be … pragmatic,” Mr Nayyar said.
Previous findings from the International Monetary Fund estimated that AI will affect 60 per cent of jobs in advanced economies and about 40 per cent of all jobs globally.
The conversation around AI's impact on the jobs market in advanced economies is fixed more on how roles might be replaced, but Mr Gaurav said it is a different story for developing nations.
“The starting point of that conversation is starkly different in developing countries, where many more people today do manual rather than cognitive work," he said.
The report said AI could boost the productivity of jobs in developing countries by 16.2 per cent, near the 18.7 per cent the World Bank expects for advanced economies.
The report listed three steps developing economies should take: adopt available tools; adapt them to local conditions; and advance towards frontier AI development. Doing so could help governments to avoid costly or inefficient mistakes if they were to try to replicate AI without putting these foundations in place.
Mr Nayyar said the UAE and Saudi Arabia were two case studies in how economies can go from adopting AI to advancing it. The nascent technology remains a key pillar for both Gulf countries' diversification agendas.
The UAE has developed the Falcon AI model, a family of large language models designed to be trained on large amounts of data and text to enhance the user's AI experience.
“You're actually finding that models that are being trained from scratch on local languages and dialects actually today are performing better than some of the other models on these local languages,” Mr Nayyar said.
But the report cautioned that building frontier AI models is not a realistic near-term goal for most developing countries.
The World Bank said that if countries in sub-Saharan Africa and elsewhere begin taking the steps to adopt AI across their economies, it can help them to achieve in a decade what would otherwise take a century.
“If they act today, if they act now, they can put into place these foundations, which will help them reap the benefits of AI down the road,” Mr Nayyar said.
