“Artificial intelligence [IA] has the potential to move workers across the distribution of skills and income. [Le défi le plus important qu’elle pose] is perhaps the potentially significant increase in inequality between workers.” A little thought on artificial intelligence and its potential impact during this back-to-school period.
The above statement is taken from the speech delivered on June 19, 2023 by Olli Rehn, Governor of the Bank of Finland, published by the Bank for International Settlements (BIS). It reminds us that this technological revolution it is generating exposes nearly 40% of jobs worldwide. In advanced countries, it will impact about 60% of jobs, the International Monetary Fund calculates. Generative AI is expected to boost productivity, revive global growth, and raise incomes worldwide. “However, it also risks replacing jobs and increasing inequality.”
Kristalina Georgieva, Managing Director of the IMF, notes that while automation and technologies have traditionally focused on repetitive tasks, AI stands out for its ability to affect highly skilled jobs. Desjardins Group economists add that unlike historical cases of the impact of disruptive technologies on the workforce, AI is more targeted at knowledge workers, who are generally more educated.
In more detail, in the so-called advanced economies, Mme Georgieva points out in her blog that, “of the jobs at risk, about half could benefit from the integration of AI, which would result in increased productivity. For the other half, AI applications could perform essential tasks currently performed by humans, which could decrease the demand for labor, lower wages and reduce hiring. In extreme cases, some of these jobs could disappear.”
In its recent projections, the analysis firm Oxford Economics estimates that depending on the degree of adoption of artificial intelligence, 12% of all jobs (or tasks) will be fully automated by 2032.
AI and the changes it brings, however, will not fail to also offer opportunities for reclassification and the creation of new jobs that we do not know exist. “Its rate of diffusion within economies and across countries, the speed at which new jobs will arrive, the social acceptability and the extent of the regulation that will sooner or later have to accompany it all remain to be determined. […]”It will be appropriate for governments to help those affected in their transition by offering training and temporary income support,” says Desjardins.
Productivity gains
The impact on productivity gains is a consensus. “AI is developing at a dizzying speed and its implementation is already having an impact on the way many companies operate. That said, the benefits on productivity may take time to really manifest themselves across the economy as a whole, but they should grow over the years,” we read in the Desjardins study.
The text published by the BIS also takes a positive view. “We find that AI, by stimulating productivity growth, significantly increases aggregate output, consumption and investment in the short and long term.” However, it adds that both higher aggregate demand and associated changes in relative prices and input costs matter much more than the direct increase in initial productivity resulting from the adoption of AI.
Oxford, for its part, has a caveat. The impact of AI and innovation on productivity may take some time. And greater productivity comes at a cost in the form of friction in the labor market.
Inflationary pressures
Beyond GDP growth, the indirect effect of generative AI will materialize on inflation and interest rates, with, as a result, an increase in productivity that will shift upwards the so-called neutral or equilibrium interest rate used as a target by central banks. “In the short term, it would add inflationary pressure, notably in the form of increased investments without immediate return or benefit. In the longer term, its impact would rather be disinflationary,” believes Oxford.
Central bankers are more hesitant, speaking of uncertain effects on inflation. On the one hand, by increasing productivity, the adoption of AI stimulates supply, which is disinflationary. On the other hand, companies must make substantial investments to take full advantage of it. This, combined with higher average incomes, will strengthen demand and increase inflation. In short, the net effect depends on the timing of these forces, which in turn depends on the expectations of households and businesses regarding the transformative potential of AI and their willingness to act accordingly.
However, in its non-deleterious version, it is said that overall, the positive contribution of AI to growth should have the effect of offsetting some adverse secular developments that threaten to depress demand in the future, including population aging, changes in global supply chains, geopolitical tensions and geoeconomic fragmentation.