Does it matter which tasks in your job AI touches, not just how many — and could that decide if you stay employed?

This explores whether the shape of AI exposure inside a job, and not only its size, decides whether people keep working. The most direct evidence in the collection says it does. An analysis of task-level AI exposure across firms from 2010 to 2023 separates two things: how much of a job AI touches on average, and how concentrated that exposure is. Higher average exposure reduces demand for labor. When exposure is concentrated in a few tasks, though, workers can shift their time to the tasks AI doesn't displace, and the net effect on employment ends up modest Does concentrated AI exposure enable workers to adapt and reallocate?. Two occupations with the same average exposure can therefore fare quite differently. A job where AI takes over one or two tasks completely leaves workers somewhere to go. A job where AI nibbles at everything leaves less room to move.

Other notes describe the same pattern in different terms. Narayanan and Kapoor argue that AI compresses the middle 'execute' layer of knowledge work, while the 'decide' and 'deliver' layers around it hold steady or grow. They point to translation and legal work, where employment has stayed stable or expanded despite AI gains Does AI really compress all layers of knowledge work equally?. That is concentrated exposure at the level of a whole profession: AI absorbs one band of tasks, and the people move to the bands around it. Acemoglu's macroeconomic model makes a similar point. Its predictions rest on tasks being reassigned between capital and labor rather than simply eliminated, which is part of why it finds AI's wage effects spread more evenly across demographic groups than earlier waves of automation. It still predicts that capital income will pull ahead of labor income Will AI increase inequality less than past automation?.

Where exposure actually lands helps show which jobs get the concentrated version. Workers have built AI into structured workflows mainly in information-intensive occupations, and this follows what the technology can do rather than older predictions about routine-task automation Where have workers actually delegated tasks to AI?. Firms also differ. Firms with more AI exposure replace online freelance workers with AI tools faster and more cheaply than other firms Do firms substitute labor for AI at different rates?. So how employment adjusts depends on the employer's own AI capability as well as on how a job's tasks are structured.

Reallocation has costs, though. The tasks workers move into are not necessarily better ones. In four experiments with 3,562 participants, people who had worked with generative AI went back to solo work with less intrinsic motivation and more boredom, because the AI had taken the engaging parts of the task and left the routine remainder Does AI collaboration drain motivation when workers return to solo tasks?. Trace data from ActivTrak show that as AI adoption rose, employees spent more time in work apps, worked more weekend hours, and had less uninterrupted focus time Does AI adoption actually reduce the work that employees do?. Concentrated exposure may protect the number of jobs while making the remaining work denser and less rewarding.

The collection has one direct study on the concentration question. The rest is supporting and adjacent evidence, so treat the 'concentration protects employment' finding as promising but not settled. If you only follow one doorway, start with the reallocation study and then read the Narayanan and Kapoor note to see the same idea applied to whole professions.