AI's reach into jobs isn't even across industries — what actually decides which sectors get hit first?

This explores how AI's reach into everyday work tasks differs from one industry to another, and what those differences mean. The corpus has no single industry-by-industry table. What it does have is a set of findings that, read together, show what drives the differences between sectors and why a simple exposure score can mislead you.

The clearest pattern is that information-heavy work is where AI has actually landed. When researchers look at tasks workers have handed off to AI inside real workflows, rather than tasks AI could in theory do, the exposure clusters in information-intensive occupations. It follows what the technology can do, not how popular chatbots are Where have workers actually delegated tasks to AI?. That contradicts the older automation story, which expected routine manual work to be hit first. This sorting also plays out across whole countries. OECD economists project AI will add 0.2 to 1.3 percentage points a year to G7 productivity growth. Most of that spread comes from each country's mix of industries: economies weighted toward AI-exposed sectors, like the US and UK, gain more than Italy or Japan How much will AI boost productivity across G7 nations?.

The less obvious finding is that the shape of exposure matters as much as how much there is. Firm-level data from 2010–2023 shows that when AI touches many tasks a little, demand for labor falls. When it hits a few tasks hard, workers move over to the tasks AI can't do, and net job effects stay modest Does concentrated AI exposure enable workers to adapt and reallocate?. Narayanan and Kapoor reach a similar conclusion from another direction. AI squeezes the 'execute' middle of knowledge work, while deciding and delivering hold steady or grow. That helps explain why translation and legal employment haven't collapsed despite heavy exposure Does AI really compress all layers of knowledge work equally?.

Who inside an industry gets exposed also varies. In male-dominated fields, exposure concentrates among high-paid, high-skill workers. In female-dominated fields, it spreads evenly across every wage level. That leaves lower-paid women exposed without the cushion high earners have Does AI exposure hit low-wage workers harder in some fields?. Even within one sector, more-exposed firms replace contract labor with AI faster and more cheaply than their peers. Adoption builds on itself inside individual companies instead of spreading evenly across an industry Do firms substitute labor for AI at different rates?.

One caveat: exposure is not the same as payoff. Getting value from AI requires reorganizing how decisions are coordinated across a company, not just swapping out tasks Does AI adoption require rethinking how organizations coordinate?. The hidden investment that requires can make heavily exposed sectors look unproductive for years before the gains appear in the statistics Why does AI productivity appear weak when investment is high?. So high exposure today may still look like a productivity slump in an industry's numbers.