The economic impact of AI: four future pathways

How artificial intelligence (AI) reshapes the economy is one of the most consequential, and most contested, questions in macroeconomics at the moment. Rather than settle on a single answer, we put bounds around it: four modelled pathways for growth, productivity, and employment. By 2060, G7 GDP could be 8% higher in an AI Breakthrough scenario, while a Disappointment scenario may see it fall by 4%.

In our central case, AI diffuses steadily rather than abruptly, raising US productivity by around 3.5% over the next decade: enough to lift growth, but slow enough that labour markets adjust without a sustained rise in unemployment. Recent financial market swings are a reminder of how wide the range around that view remains. We model three alternatives: Breakthrough, Disruption, and Disappointment, each varying different assumptions about AI adoption, productivity gains, and labour market outcomes based on our new AI scenario levers. Breakthrough pairs rapid adoption with labour markets flexible enough to absorb the shift. Disruption delivers comparable productivity gains into rigid labour markets, generating structural unemployment. Disappointment sees adoption and productivity gains fall short of what markets have already priced in.

AI’s gains will not be evenly shared. The G7, most exposed to the technology, stands to gain the most; economies with weaker digital connectivity, such as India, considerably less. Running every scenario through our Global Economic Model lets us quantify that divergence in full, across every economy we cover and every indicator, asset returns included.

Download the report for more detailed insights.