Need to assess economic implications of artificial intelligence

Ranjit Singh Ghuman

Professor of Eminence, Guru Nanak Dev University

It is being argued that artificial intelligence (AI) would enhance the productivity of labour and capital and hence push economic growth and the GDP, pulling the global economy out of the slowdown. But AI would have socio-cultural and politico-economic implications. The large-scale use of AI would lead to a massive displacement of labour and thereby result in a net loss in employment. This, in turn, would require massive financial resources to reskill the displaced workers and take care of their financial requirements. The provision of universal basic income (UBI) is being projected as one of the solutions. However, there are many ifs and buts about whether the beneficiaries of AI and enhanced productivity would spare funds for implementing the UBI and reskilling the displaced workers.

In the 1960s, economists argued that a high growth rate and a large size of GDP would result in trickle-down benefits from a higher growth rate and hence lower economic inequality. But it did not happen in reality, as income and wealth inequality have been increasing, leading to a slowdown in the demand for goods and services and a decline in the demand for labour.

The average annual growth rate of the world economy has declined from 3.2 per cent during 1980-2010 to 2.7 per cent during 2010-2020, and the per capita income growth rate declined from 1.69 per cent to 1.54 per cent per annum (World Bank’s World Development Indicators 2023). The World Bank’s January 2024 report (Global Economic Prospects) forecasts that “most economies — advanced as well as developing — are set to grow slower in 2024 and 2025 than they did in the decade before Covid-19. Global growth is expected to slow down for a third year in a row — to 2.4 per cent—before rising up to 2.7 per cent in 2025. Those rates, however, would still be far below the 3.1 per cent average of the 2010s. The per-capita investment growth in 2023 and 2024 is expected to average just 3.7 per cent — barely half the average of the previous two decades. Without corrective actions, global growth will remain well below its potential for the remainder of the 2020s.” The report also predicts a significant decline in the income of people across most countries.

The average annual growth rate of the value of the human development index witnessed a sharp decline from 0.78 per cent during 2000-2010 to 0.45 per cent during 2010-2020 (UNDP: Human Development Report, 2021-22). The global employment elasticity during 2010-2019 was minus 0.85 per cent, which means a one percentage point increase in the growth rate is leading to a 0.85 percentage point decline in employment. Employment elasticity across agriculture, industry and services was minus 0.92 per cent, minus 0.78 per cent and minus 0.88 per cent, respectively, during the same period. It is evident that both the aggregate and sectoral growth rates are generating a net loss in jobs.

The IMF’s recent study, ‘Artificial Intelligence and the Future of Work’, reveals that AI will affect 40 per cent of all jobs around the world, while in the advanced economies, it may impact 60 per cent jobs. Its effect on the emerging markets and low-income countries may be confined to 40 per cent and 26 per cent, respectively, in the near immediate future. At the same time, many of these countries may face an uphill task to restructure and reskill the displaced workforce, as they lack the required physical infrastructure and financial resources.

The report further says that though productivity will rise with the adoption of AI, it will also increase income and wealth inequality across countries.

Unemployment and inequality have generally increased or remained stubbornly high worldwide over a span of three decades (1991-2019). The proportion of vulnerable employment and that of the working poor has also increased. All this has resulted in shrinking the purchasing power of a large proportion of the population. This provides a plausible explanation for a global slowdown and the low aggregate demand for goods and services, and thereby for the ongoing recession for about two decades. Such a scenario is likely to pose a serious threat to the sustainability of the market-driven consumption-led development paradigm and the feasibility of sustainable development goals.

Abhijit Banerjee and Esther Duflo (winners of the Nobel Prize in economics), in their 2019 book, Good Economics for Hard Times, raised serious concerns about the socio-economic implications of AI-led technology, the displacement of human labour with robots and the UBI. Let us suppose for a while that the UBI may take care of the financial requirements of the displaced workers, which has a very low probability. The issue is: How will it take care of the self-esteem and dignity of the displaced and jobless workers and people living on ‘charity’? Will such people, especially the youth, not develop a sense of neglect, irrelevance and inferiority? And how would they react? What would be our future social structure and demographic composition? Will the beneficiaries of higher productivity and profits driven by robot-driven technology be really willing (and for how long?) to compensate the displaced workers and provide finances to support and sustain the UBI? How would the system deal with the consequent high inequality and the shrinking purchasing power of the people? Will it not be a drag on the aggregate effective demand for goods and consumption and thereby on the sustainability of the market-driven, consumption-led development paradigm? Will such a scenario not pose a serious threat to socio-cultural and political-economic stability?

This, however, does not mean that AI-driven technology or automation should not be developed. It is going to be there. But the basic question is: How to deal with the socio-cultural and politico-economic implications of AI and UBI? All this requires an informed public debate.

Join Whatsapp Channel of The Tribune for latest updates.

#Artificial Intelligence AI