Image: Getty Images
A MACHINE THAT can think, or at least appear to, is not the stuff of science fiction anymore. The arrival of “generative” artificial intelligence, the sort that powers ChatGPT, marks the dawn of a new era of automation. For South Asia, a region of nearly 2bn people, this presents both a grave threat and a glittering opportunity. Artificial intelligence could eviscerate industries that provided millions with a ladder into the global middle class. It could also supercharge development, bypassing the need for some of the grinding industrialisation of the past. The outcome depends on whether governments can prepare their people and economies for what is to come.
Technological change has always been disruptive. In 1860 more than half of American workers were employed in agriculture. Today the figure is about 1%. Yet the transition, though often painful for individuals, was manageable for society. New jobs were created even as old ones were destroyed. Some 60% of American workers today are in jobs that did not exist in 1940. The computing revolution of recent decades eliminated many clerical and administrative roles, but created many more for designers, developers, managers and so on.
Artificial intelligence is different. It is a “general-purpose technology”, like the steam engine or electricity, which could transform swathes of the economy. It is spreading faster than previous waves of technological change, in part because it requires little physical investment beyond a computer and an internet connection. And its effects will be felt most in white-collar office work, not on the factory floor. It threatens mid-skilled cognitive jobs, the sort of business-processing work that has been a source of growth and employment in South Asia.
The region’s back offices are vulnerable. Since the launch of ChatGPT in late 2022, research has shown that wages and hiring have fallen in sectors with high exposure to the new technology. The worst-affected roles are those where artificial intelligence can easily substitute for human workers, rather than complement them. In business services, such as marketing, finance and design, artificial intelligence is already having a noticeable effect. Mid-skilled and entry-level workers are most at risk.
The new technology is not all bad news for South Asia. Some high-end digital-service exports, such as computer programming, have risen since the launch of ChatGPT, suggesting that artificial intelligence can be a useful tool for the well-trained. But the region is not well prepared to harness these benefits. South Asia scores poorly on the International Monetary Fund’s “AI Preparedness Index”, which measures digital infrastructure, human capital, innovation and regulation. Only India is near the average for emerging markets. Its neighbours lag behind.
Start with the basics. Access to electricity is nearly universal across South Asia, but the quality and speed of broadband internet are poor. Fixed-line broadband speeds are roughly a quarter of those in the rich world. Some 60% of South Asians use the internet, the lowest rate of any emerging region. The urban-rural divide is stark. Just 36% of rural South Asians are online, the widest such gap in the emerging world. Without better and more inclusive internet access, the benefits of artificial intelligence will be limited to cities and the wealthy.
Then there is education. South Asia has many excellent universities and produces world-class computer scientists. Too many of them leave. The majority of India’s best artificial-intelligence researchers emigrate to America or Europe. For those who stay, the quality of education tends to be poor. Some 75% of South Asians are literate, below the average of 85% for emerging markets. Artificial intelligence could help, by providing personalised tutoring or adapting lessons to the level of a student. But that needs children to be in school, and for schools to be equipped with the necessary technology.
Even if the people and infrastructure were ready, South Asia’s business environment may not be. The region’s firms are mostly small, with limited access to credit. They tend to be slow to adopt new technology. Artificial intelligence is concentrated in big companies. The region’s “start-up” culture is vibrant in places, but it is no Silicon Valley. The world’s artificial-intelligence hubs are built on vast pools of capital, top-notch talent and a tolerance for failure. South Asia’s innovators face limited funding, onerous regulation and weak intellectual-property protection.
For governments, therefore, the task is twofold. They must manage the disruption that artificial intelligence will cause to those whose jobs are at risk. They must also try to maximise the benefits that the new technology can bring. That means investing in digital infrastructure and education. It means freeing markets, so that capital and ideas can flow to the most promising ventures. And it means using artificial intelligence to improve the business of government itself, by cutting corruption and improving public services.
The history of modern South Asia is, in part, a story of using technology to leapfrog stages of development. Mobile phones allowed the region to skip the construction of telephone wires. Digital payments are reducing the dependence on cash. Artificial intelligence offers a similar opportunity. It could allow the region to bypass some of the drudgery of industrialisation, and create a new development path. The alternative is a new sort of dependency: on technology developed elsewhere, operated by a small elite, with the gains flowing across the ocean. The thinking machines are here. South Asia must now learn how to put them to work. ■
