As a writer, I use generative AI almost every day to help research stories, test ideas, challenge arguments and work through drafts. It has changed the way I work and, no doubt about it, how much I can accomplish in a given amount of time.
I’m hardly alone. Millions of people are using generative AI in their work to write computer code, analyze information, answer customers, prepare presentations, brainstorm ideas and perform countless other tasks.
That raises a larger question: As AI changes how much work one person can accomplish, what is it doing to jobs?
AI moves into the workplace
The shift to AI-assisted, and in some cases AI-driven, work is already well underway. More than half of US workers now say they use AI in their jobs, according to Gallup. Thirty percent use it at least a few times a week, and 15 percent use it every day. Nearly half also say their employers have integrated AI tools into their organizations.
Generative AI in particular has spread widely. From August 2025 through May 2026, Federal Reserve researchers surveyed nearly 14,000 workers and found that at least 20 percent of workers used gen AI in more than 80 percent of occupations. Usage was especially high in information technology and business and financial operations.
What are workers doing with it? Much of it looks surprisingly familiar. They are writing and editing, searching for information, analyzing documents, generating ideas and solving problems. Gallup found that chatbots and virtual assistants were the most commonly used AI tools, followed by writing and editing tools.
So it’s not a news flash that AI is already becoming part of everyday work. But is it really helping people do work better, or at least faster?
AI changes how much one person can do
Researchers are finding that workers can accomplish certain kinds of work faster, and sometimes better, with AI’s help.
One of the most revealing studies took place at Procter & Gamble. Researchers divided 776 experienced professionals working on real product-development challenges into individuals and two-person teams, with some given access to gen AI and others working without it.
The results were striking. Individuals working with AI produced work comparable in quality to two-person teams working without AI. The researchers also found that AI helped employees work beyond their usual areas of expertise. A research and development specialist, for example, could contribute effectively to the going-to-market side of a problem, and vice versa.
That doesn’t mean P&G found it could replace two employees with one. But it demonstrated that under the right circumstances, AI can seriously increase what one person can accomplish.
Other research points in the same direction. In a study involving more than 5,000 customer support agents, workers given an AI assistant resolved about 14 percent more customer problems per hour. The gains were much larger among less experienced and lower-skilled workers, who improved by roughly a third.
When AI productivity cuts both ways
The implications for jobs are more complicated. This website is one small example. I’m a writer, not a web developer, yet with AI’s help, I was able to build almost all of Outshining AI myself.
That could represent work I might once have hired someone else to do. But more likely, without AI, I would never have created the site at all. ChatGPT didn’t just help me do existing work more efficiently. It enabled me to create something new.
Multiply both possibilities across millions of workers and businesses, and the employment picture becomes much harder to predict. AI can reduce the labor needed to accomplish something while also increasing the number of things people are able to create.
That means AI-driven job loss won’t necessarily announce itself through mass layoffs. It can also appear quietly: a position that isn’t filled, a job that isn’t created or outside help that is no longer needed. At the same time, greater productivity can help businesses expand and create work that didn’t exist before.
Where AI job loss is showing up
So what does the evidence look like on the ground? So far, the clearest effects are concentrated rather than widespread. They are appearing more in some occupations than others, and more among people entering those fields than among experienced workers.
- Software development: Young software developers have been among the hardest hit. Stanford researchers found substantial employment declines among 22-to-25-year-olds in software development, even as employment among older workers held up much better. Job postings have rebounded since May 2025, according to Indeed, but most of the growth has been in senior positions. More than a third of that growth came from jobs with AI in the title.
Customer service: The pattern is similar. Stanford found declining employment among young customer-service workers, another occupation highly exposed to generative AI. This is also the field in which the study cited earlier found that an AI assistant allowed workers to resolve about 14 percent more customer problems per hour. Customer service job trends bear close watching. - Clerical and other knowledge work: The picture is less dramatic here. Stanford has found early-career employment weakness in AI-exposed clerical work, but broad displacement across professional occupations has not happened. Other research has found AI changing the tasks people perform without yet producing large changes in their hours or earnings.
- Young workers across AI-exposed occupations: The broader pattern is especially striking. Stanford found that employment among 22-to-25-year-olds in highly AI-exposed occupations was 19 percent below where it would have been if it had kept pace with their peers in less-exposed occupations. Experienced workers showed no comparable gap. The difference appears to be driven mainly by reduced hiring rather than layoffs.
The evidence so far points to an early divide. Experienced workers in AI-exposed fields appear to be holding their own, while younger workers trying to enter some of those same fields are finding fewer job opportunities. AI’s clearest employment effect may not be mass layoffs. It may be companies hiring fewer entry-level workers as AI-equipped employees can accomplish more of the work themselves.