Goldman Sachs estimates that AI could displace up to 9% of workers, or roughly one in 11. This means technology could replace them in their current jobs. Economist Joseph Briggs bases the estimate on a forecast that AI will raise productivity—the amount produced in a given working time—by 15% in the US and other developed economies. Slower recruitment could reduce employment by a further 3%.
The timing depends partly on the economy. Briggs found that technology-related job losses tend to increase during recessions, when economic activity falls. Companies often use these periods to replace tasks with machines and reduce staff costs. However, he expects the displacement to be temporary. His estimate measures jobs lost, rather than unemployment: some affected workers may quickly find another position, while others may leave the workforce.
So far, the effects have been smaller than people feared earlier in 2026. Only 23% of US workplaces have introduced AI into their regular activities, and most expected productivity improvements are still ahead. Meanwhile, the economy and job market have remained strong, giving employers less reason to cut labour costs aggressively. Briggs warns that a downturn could speed up AI-related job losses and make the economic slowdown worse.
