McKinsey says AI could create more jobs than it destroys, but millions still face displacement

The AI-will-steal-your-job narrative just got a major counterweight. McKinsey Global Institute released a report projecting that artificial intelligence may ultimately generate more employment than it eliminates, though the path from here to there won’t be painless for millions of American workers.

The report, titled “Agents, robots, and us: Skill partnerships in the age of AI,” paints a picture of a labor market in transformation rather than freefall. The core finding: roughly 57% of current US work hours have the theoretical potential for automation, but that number represents a ceiling, not a forecast of actual job losses.

The $2.9 trillion collaboration thesis

McKinsey’s central argument rests on the idea that the biggest economic gains won’t come from replacing humans with machines. They’ll come from redesigning workflows so humans and AI work together.

The projected payoff for getting that collaboration right: up to $2.9 trillion in additional annual value by 2030. One of the more reassuring data points in the report: over 70% of skills that employers currently seek are applicable to both automatable and non-automatable tasks. In practical terms, that means most workers aren’t facing a binary outcome where their entire skill set becomes worthless overnight.

The gap between fear and reality

McKinsey’s analysis builds on research the firm first published in 2017, which argued that sufficient investment in workforce transitions could lead to net job growth even in an era of aggressive automation.

Subsequent surveys cited in the research found that AI-related workforce reductions amounted to roughly 14%, significantly lower than the 32% that had been anticipated. That’s a meaningful gap between what people feared and what actually materialized, suggesting that the technical capability to automate a task and the economic incentive to actually do so are two very different things.

AI fluency is the new literacy

Perhaps the most striking trend in the report is the velocity of change in skill demand. The demand for AI fluency has surged approximately sevenfold over the past two years, according to McKinsey’s analysis.

The report envisions a future workplace organized around what it calls “skill partnerships,” where humans handle tasks requiring judgment, creativity, and interpersonal nuance, while AI agents manage cognitive workloads and robots take on physical tasks.

What this means for markets and workers

The report also suggests that companies investing early in human-AI workflow redesign will have a structural advantage over competitors who treat AI adoption as a simple headcount reduction exercise.

But the transition costs are real and unevenly distributed. Workers in roles with high automation potential, particularly routine cognitive tasks like data entry, basic analysis, and standardized reporting, face the most immediate pressure.

The 57% automation-potential figure is a useful benchmark precisely because it highlights both the scale of what’s possible and the gap between possibility and implementation. Actual displacement has tracked far below theoretical maximums, with AI-related workforce reductions coming in at 14% against an anticipated 32%.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.