Does AI help beginners and experts in the same way at work, or does each group trade something different for the boost?
Does AI assistance reduce effort differently for novice versus expert workers?
This explores whether AI makes work easier in different ways for beginners and for experienced workers, and what each group gives up or gains in exchange.
This explores whether AI makes work easier in different ways for beginners and for experienced workers. One caveat comes first: the collection measures output, quality and time much more often than it measures effort itself. Even so, a clear and slightly uncomfortable pattern shows up. Novices gain the most in the moment, and they may also pay the most later.
The clearest evidence that novices gain most comes from a study of more than 5,000 customer support agents at a large firm. AI assistance raised the number of issues resolved per hour by 15% on average, and nearly all of that gain went to less experienced agents, who got both faster and better. The most experienced agents barely sped up, and the quality of their work dipped slightly (Does AI assistance help less experienced workers most?). One reading is that the AI was spreading the know-how of top performers to newcomers, while giving experts suggestions they didn't need and that sometimes pulled them off course. That links to a separate finding: an AI suggestion can break someone's concentration even when it's correct, and the person then has to rebuild their focus (Does AI assistance always help reasoning or does it carry hidden costs?). For an expert who is deep in their work, help can cost more than it saves.
The novice advantage has a catch. Productivity gains tend to appear when people apply skills they already have. When workers use AI to learn something new, the gains disappear and their learning suffers (When does AI actually boost worker productivity?). Workers who did much better with generative AI showed no improvement when they later did similar tasks on their own (Does AI assistance help workers learn lasting skills?). So the effort AI saves a novice may be exactly the effort that would have turned them into an expert. Confidence adds to the problem. In one survey, 90% of workers said they felt confident with AI, yet only 25% said it worked on the first try, and half had spent more time with AI than doing the task by hand. The gap was widest among younger workers (Why do workers feel confident with AI but get poor results?).
For experts, the risk is different. AI doesn't so much fail to help them as slowly wear down the skills that make them experts. Anthropic engineers reported large productivity gains, yet most said they could fully hand off only a small share of their work. They worried that leaning on the AI for routine coding would erode the hands-on practice they need to catch its mistakes (Does AI assistance erode the skills needed to oversee it?). Experts can supervise AI well because of their experience, and that experience comes from doing the routine work AI now absorbs.
The idea that AI "reduces effort" may also be the wrong frame. One line of research finds that AI doesn't cut total task time. Instead, it moves time away from doing the work and toward writing prompts and checking what comes back (Does AI really save time, or just change how we spend it?). Effort changes shape rather than disappearing, and checking output is far easier for someone who already knows what good work looks like. Something also happens to motivation: after working with AI, people felt more in control but less motivated and more bored when they returned to solo work, because the AI had taken over the engaging parts of the task (Does AI collaboration drain motivation when workers return to solo tasks?). At a larger scale, AI seems to compress the "doing" layer of knowledge work while deciding and delivering stay the same or grow (Does AI really compress all layers of knowledge work equally?). Experienced workers spend more of their time in those layers. The open question the collection doesn't yet answer is how today's novices will become the experts who can do that judgment work if AI handles the practice that used to train them.
Sources 9 notes
A study of 5,172 support agents at a Fortune 500 firm found a 15% average productivity gain from AI assistance, with gains concentrated among less experienced workers who improved both speed and quality. The most experienced agents saw small speed gains but slight quality declines.
Well-intentioned AI suggestions can damage reasoning performance by severing cognitive immersion, forcing users to rebuild focus before continuing. Evaluation must measure flow preservation across entire tasks, not just local suggestion accuracy.
Studies showing AI productivity gains measured tasks within workers' existing domains. When workers used AI to learn new skills, productivity gains disappeared and learning suffered, suggesting prior findings do not generalize to skill acquisition.
Wu et al. found that workers using generative AI performed substantially better on content tasks, but when performing similar tasks independently afterward, their performance showed no improvement. The capability did not transfer across contexts.
WalkMe's survey of 2,037 US workers found 90% feel confident using AI, but only 25% report it works on first try and 50% spent more time using AI than doing tasks manually. The gap widened most among younger workers, suggesting overestimation of skill.
Show all 9 sources
Anthropic's 132-person survey found 50% self-reported productivity gains and 67% more merged pull requests, yet most engineers can only fully delegate 0-20% of work. Employees fear that relying on Claude for routine tasks erodes the hands-on coding practice needed to catch its errors.
Research shows AI doesn't reduce total task time; it reallocates it away from active work toward composing prompts and understanding outputs. This shift changes the cognitive demands and learning outcomes, making time-on-task a poor productivity metric.
Four experiments (N=3,562) found that after collaborating with GenAI, workers gained sense of control in solo work but experienced lower intrinsic motivation and higher boredom. AI had absorbed the engaging parts of tasks, leaving mundane residual work.
Narayanan and Kapoor argue AI narrows only the middle execution layer of knowledge work while decide and deliver layers persist or grow. Translation and legal work show stable or expanding employment despite AI gains, suggesting task-level compression doesn't shrink occupational demand.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- How AI Impacts Skill Formation
- Does generative AI narrow education-based productivity gaps? Evidence from a randomized experiment
- Toward Measuring AI's Effects on Skill Formation: The Stock-Formation Gap
- Adoption of Generative AI in the Workplace: Increasing and Shifting the Balance of Productivity and Communication Activity
- What 81,000 people told us about the economics of AI
- Research: Gen AI Makes People More Productive—and Less Motivated
- Generative AI at Work
- Beyond Productivity: Measuring the Real Value of AI