Does AI help the least skilled workers most, or does it mostly help when they're applying skills they already have, not learning new ones?

This explores whether AI narrows the gap between less-skilled and more-skilled workers, giving the biggest lift to those starting furthest behind. The collection has no head-to-head study that measures productivity gains by worker ability, so it can't give a clean yes or no. What it does have is a set of findings from different angles, and together they suggest the question needs a qualifier: gains for which kind of 'low ability'?

The sharpest qualifier comes from looking at what earlier productivity studies actually measured. The gains they reported came from workers doing tasks inside their own field. When people used AI to learn a skill they didn't already have, the productivity gains disappeared and their learning suffered When does AI actually boost worker productivity?. So 'low ability' can mean two different things. A junior person in a familiar role may be lifted toward expert output. Someone who lacks the underlying skill entirely gets much less, because AI seems to amplify competence rather than stand in for it. Even the lift can carry a cost: research mapping thousands of workplace AI scenarios warns that relying on AI can slowly wear down the skills and judgment workers need to supervise it Does AI augmentation protect workers from skill erosion?. Over time, a short-term gain for a novice could turn into a ceiling.

One result does point toward equalizing, though in hiring rather than productivity. In a hiring experiment, listing AI skills partly or fully offset the interview penalties faced by older candidates and those without a bachelor's degree. The effect was strongest for office assistant roles Can AI skills help older or less-educated job candidates?. There is a social catch, however. People who use AI expect to be judged as less competent and less diligent, so they often hide it from managers Do people fear judgment when they use AI at work?. Workers whose competence is already questioned may pay the steepest price for that perception.

Exposure to AI is also distributed unevenly, which complicates who gains. In male-dominated occupations, exposure clusters among high-skilled, high-paid workers. In female-dominated occupations it spreads evenly across skill levels, which leaves lower-paid women exposed with fewer resources to adapt Does AI exposure hit low-wage workers harder in some fields?. Data on tasks workers have actually handed over to AI shows delegation concentrated in information-heavy work, with wage patterns reversing at the advanced-degree level Where have workers actually delegated tasks to AI?. Whether exposure turns into gain or loss depends partly on how much of a job AI touches. When only a few tasks are affected, workers can shift toward the rest Does concentrated AI exposure enable workers to adapt and reallocate?.

The question also changes at a larger scale. Anthropic's scenario model suggests that even if AI compresses skill gaps within a team, the bigger shift may run from workers as a whole toward capital owners, with knowledge-worker wages stagnating as GDP grows Does AI growth inevitably shift wealth away from workers?. Firms that are already heavy AI users replace freelance labor faster and more cheaply than other firms do Do firms substitute labor for AI at different rates?. Narrowing the gap between workers can happen alongside, and partly hide, a widening gap between workers and the firms that own the AI.