Dark Leisure

Many experts are questioning whether we are in an AI “bubble,” as impressive investments and predictions about artificial intelligence have yet to translate into clear productivity gains or widespread worker displacement. Megan McArdle draws parallels to previous technology trends, like computers and the internet, whose full economic impact took time to appear in statistics. She observes that while companies are spending billions on AI infrastructure, the promised revolution is still unfolding slowly, with gains masked by factors like "dark leisure" with workers using AI to finish tasks faster and spending extra time on non-work activities. True productivity benefits may be delayed as managers reset expectations and as society adapts to new workflows.

McArdle argues that part of the explanation lies in the slow learning curve for both workers and organizations. While advanced users may see incremental benefits, the larger transformation will come as new work styles emerge and AI enables people with different skill sets to do new kinds of work. The history of earlier technologies suggests that, over time, AI will drive major efficiency gains throughout the economy.

A counterfactual scenario would ask: What if the hype about artificial intelligence fails to deliver at scale, and productivity never meaningfully improves? In this version of events, businesses might face substantial losses as investments in AI infrastructure fail to yield returns, leading to reduced enthusiasm for AI, layoffs in related sectors, and skepticism toward future tech innovations. Workers would not see significant improvements in job automation or efficiency, and policymakers might shift focus back to traditional growth drivers rather than digital transformation. The expectation of creative and radical job redesign would ultimately be replaced by disappointment and recalibration of technological bets.