The impact of AI on the economy: potential and challenges for global growth

I'm Alessia Gavazzi, Economist, Intesa Sanpaolo.

I'm Mario di Marcantonio, US Economist, Intesa Sanpaolo
 

Alessia Gavazzi

Artificial intelligence represents a technological breakthrough comparable, in terms of scope and pervasiveness, to the advent of electricity and the internet. It redefines the boundary between tasks that can be automated and activities with high human added value. The figures confirm the scale of the phenomenon: in the United States, the AI-related technology sector contributed 45% to GDP growth recorded in 2025, more than double the 20% contribution observed during the dot-com era¹.

Although the spread of AI is rapid, actual adoption within production processes is proceeding at a slower pace than surveys suggest. Whilst a quarter of global organisations report adopting agentic AI systems – that is, AI that operates autonomously to manage complex tasks – penetration rates across G7 economies still show significant discrepancies, ranging from 2% in Italy to 6% in the United States². On the investment front, around $3 trillion has been announced over the next five years³ for infrastructure and data centres. The main obstacle is the energy constraint, which could put upward pressure on prices and act as a bottleneck for the infrastructure to become fully operational.

Mario Di Marcantonio

It is estimated that the impact of Artificial Intelligence on global productivity will range between 0.1% and 1.5% per annum over a ten-year horizon. However, the impact will not be linear. The most plausible trajectory is that of a J-curve, i.e. an initially modest impact, followed by an acceleration once adoption is accompanied by investment in human capital and process reorganisation. In the European Union, stricter AI legislation constitutes a potential constraint on the pace of growth.

Regarding employment, the prevailing view is that AI will transform work rather than replace it. The advent of AI acts as a catalyst for the search for sophisticated skills, thereby increasing the demand for more highly qualified professionals. One cause for concern relates to the recruitment of entry-level staff, which has fallen by 16%⁴ in the sectors at risk. For countries such as Italy, where the workforce is shrinking due to an ageing population, AI could partially offset the decline in GDP caused by demographic contraction, thanks to increased productivity.

In conclusion, the benefits of AI will depend on the ability to integrate technology, infrastructure and human capital. For Europe, the challenge is to prevent the competitive lag from turning into a structural divide. Managing the transition will be crucial to ensuring that technology becomes a driver of growth.

1 The contribution is given by the sum of the components: research and development, software, intellectual property products, computers and peripheral equipment, and information processing equipment. Source: Intesa Sanpaolo calculations based on data from the Federal Reserve Bank of St. Louis and the U.S. Census Bureau.

2 SOURCE: https://www.oecd.org/en/publications/miracle-or-myth-assessing-the-macroeconomic-productivity-gains-from-artificial-intelligence_b524a072-en.html

3 SOURCE: White House, https://www.whitehouse.gov/investments/

4 SOURCE: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai