Will most worker skills survive AI automation?

McKinsey Global Institute's report argues that although today's technologies could in theory automate activities accounting for "about 57 percent of US work hours," most of the skills people currently use will not become obsolete — the report finds "more than 70 percent of today's skills can be applied in both automatable and non-automatable work." MGI frames this as a shift in how skills are used rather than a disappearance of them: workers will "spend less time preparing documents and doing basic research... and more time framing questions and interpreting results." Alongside this, MGI reports that "demand for AI fluency — the ability to use and manage AI tools — has jumped nearly sevenfold in two years," while "job post mentions are declining for routine writing and research."

MGI built these estimates from a new Skill Change Index (SCI), "a time-weighted measure of automation's potential impact on each skill used in today's workforce," derived from Lightcast job-posting data covering "roughly 6,800 skills cited frequently in more than 11 million job postings." The 57 percent automation-potential figure "compares the capabilities of existing technologies, including those demonstrated in a lab, with the level of human proficiency required for different work tasks" — a technical-potential estimate, not a job-loss forecast, and one the report says "should be updated regularly" as technology advances. MGI ties the skill-overlap finding to its broader argument that capturing the report's estimated "$2.9 trillion" in annual US economic value by 2030 depends less on new AI breakthroughs than on "how organizations redesign workflows" and "how quickly human skills adapt."

This complements Is AI creating common skills across jobs or deepening divisions?, which found AI-related vacancy demand converging on a narrow technical core (Python, SQL, machine learning) — MGI's sevenfold rise in AI-fluency postings is consistent with that convergence, though MGI frames AI fluency as a widely distributed complementary skill rather than a specialist core. It also sits alongside What makes accountable judgment scarce when AI cognition is cheap?: where that insight locates the scarce asset in judgment and institutional accountability, MGI locates it in the breadth of skills that transfer across automatable and non-automatable work, without specifying which institutions determine that transfer. It extends Does automation raise or lower the skills that remaining work demands? by supplying a concrete mechanism — the SCI — for tracking which skills rise or fall in demand as tasks are automated.

The 57 percent and 70 percent figures are, by MGI's own description, technical-potential estimates built from lab-demonstrated capability compared against job-posting language, not measurements of actual displacement, adoption speed, or which specific workers will be affected; the report explicitly declines to forecast job losses. The job-posting data also "reflect hiring intentions rather than the actual work people do," so the AI-fluency and skill-decline figures describe what employers say they want, not confirmed shifts in day-to-day task performance. The analysis is further scoped to the United States, with the report noting only that its patterns "apply broadly" elsewhere rather than demonstrating this.