What Has Generative AI Actually Done to Creative Jobs? The Evidence Through 2025
Doug Woodham | September 2026
U.S. employment data through 2025 show no broad collapse in creative jobs. But writers, artists, and graphic designers lost ground while other creative occupations grew, revealing where generative AI is reshaping work
Generative AI has advanced with extraordinary speed. Since 2023, tools that can produce text, images, and music have become far more capable and widely used. Predictions quickly followed that creative workers would be among the first to face substantial job displacement as AI took over more of their work.
Until recently, there was little labor-market evidence with which to judge those claims. There is now enough data to begin doing so. I use 2023 as the baseline because it predates the widespread adoption of generative AI in creative work. The comparison therefore asks what changed as these technologies became more capable and more widely used during 2024 and 2025.
The headline is surprising. Across the 15 creative occupations tracked in this analysis, employment was virtually unchanged between 2023 and 2025, declining by less than 1 percent. Beneath that stability, however, occupations moved in sharply different directions. Employment fell among writers, graphic designers and artists, yet rose substantially among architects, producers and directors. In some occupations, the larger deterioration appeared not in headcount but in a shift from full-time to part-time work or in weaker relative earnings.
This is neither mass displacement nor business as usual. Creative labor markets are diverging. The question is why some occupations have weakened while others, facing a similar technological shock, have grown.
The sections that follow first establish the size and structure of the creative workforce, then examine where employment, full-time work and pay strengthened or weakened between 2023 and 2025. The final sections turn to the harder question: why exposure to the same technology appears to have produced such different labor-market outcomes by occupation.
What Counts as a Creative Job?
This analysis tracks 15 occupations in which creating or shaping images, language, performances or designs is central to the work. The classification is based on occupation rather than industry. A graphic designer working for a bank belongs in the creative workforce; an accountant working for a film studio does not. The analysis draws on data from the U.S. Bureau of Labor Statistics; details on the data sources, occupational definitions and methodology appear in the appendix.
Exhibit 1: Design is the largest part of America’s creative workforce
Employment in 15 creative occupations, United States, 2023
| Broad area / occupation | 2023 employment |
|---|---|
| Design | 901,000 |
| Other designers | 384,000 |
| Graphic designers | 312,000 |
| Interior designers | 103,000 |
| Floral designers | 63,000 |
| Fashion designers | 39,000 |
| Performing arts | 449,000 |
| Musicians and singers | 170,000 |
| Producers and directors | 167,000 |
| Actors | 50,000 |
| Music directors and composers | 46,000 |
| Dancers and choreographers | 16,000 |
| Writing | 396,000 |
| Writers and authors | 271,000 |
| Editors | 125,000 |
| Artists and related workers | 334,000 |
| Photographers | 228,000 |
| Architects | 203,000 |
| Total creative workforce | 2,511,000 |
Design was by far the largest group, with just over 900,000 workers shaping everything from the graphics on our screens to the interiors of our homes and the visual identities of the brands around us. The largest occupation within it was the somewhat opaque Bureau of Labor Statistics category “Other Designers.” Its 384,000 workers include exhibition designers, visual merchandisers and creators of themed environments, among other specialized roles.
The performing arts accounted for another 449,000 workers, from producers and directors who shepherd projects from conception to completion to musicians, actors and dancers who bring them to life. Writing employed nearly 400,000 people. Artists, photographers and architects rounded out the workforce.
Together, the 15 occupations employed just over 2.5 million people in 2023. That was roughly 600,000 more than worked in legal occupations, making the creative workforce about one-third larger than the legal workforce.
Headcount alone does not tell us how strong a labor market is. Two other measures matter throughout this analysis: whether work is full time or part time, and how pay compares with the broader labor market. Across the creative occupations, both varied sharply at the outset.
Exhibit 2: Creative jobs started from very different positions
United States, 2023
Full-time work varies widely
Share of people at work who work full time, %
| Actors | 51 |
|---|---|
| Musicians & singers | 56 |
| Floral designers | 61 |
| Photographers | 65 |
| Dancers/ |
67 |
| Fashion designers | 71 |
| Music directors/ |
73 |
| Writers & authors | 73 |
| Artists | 73 |
| Interior designers | 74 |
| Editors | 78 |
| Graphic designers | 87 |
| Other designers | 87 |
| Producers & directors | 92 |
| Architects | 92 |
Most creative jobs pay above the national median
Median earnings of full-time wage-and-salary workers, national median = 100
| Actors | 110 |
|---|---|
| Musicians & singers | 122 |
| Floral designers | 64 |
| Photographers | 102 |
| Dancers/ |
111 |
| Fashion designers | 121 |
| Music directors/ |
93 |
| Writers & authors | 118 |
| Artists | 138 |
| Interior designers | 112 |
| Editors | 129 |
| Graphic designers | 112 |
| Other designers | 133 |
| Producers & directors | 131 |
| Architects | 164 |
Some occupations were overwhelmingly full time; others relied much more heavily on part-time work. Pay varied too, but most creative occupations began from a relatively strong position. Occupations whose median full-time earnings met or exceeded the national median of $58,084 accounted for roughly 96 percent of the creative workforce.
These measures matters because labor-market deterioration can take several forms. An occupation can lose workers, lose full-time jobs to part-time work without much change in total employment, or lose ground in pay relative to the rest of the economy.
Creative Occupations Moved in Very Different Directions
From 2023 to 2025, creative employment barely moved in aggregate. The occupations inside it did.
Exhibit 3: Creative occupations moved in sharply different directions
Change in employment, United States, 2023–2025, percent
| Floral designers | -22 |
|---|---|
| Writers and authors | -14 |
| Actors | -14 |
| Artists and related workers | -10 |
| Editors | -10 |
| Graphic designers | -9 |
| Photographers | -5 |
| Musicians and singers | -2 |
| Other designers | -1 |
| Fashion designers | +13 |
| Dancers and choreographers | +13 |
| Interior designers | +21 |
| Architects | +25 |
| Music directors and composers | +26 |
| Producers and directors | +29 |
Employment among writers and authors fell 14 percent, artists and related workers 10 percent, and graphic designers 9 percent. At the other end of the chart, interior designers rose 21 percent, architects 25 percent, and producers and directors 29 percent.
The larger point is clear: there was no common direction of travel across creative work. A virtually unchanged aggregate concealed labor markets moving sharply apart.
Creative Job Losses Were Only Part of the Story
Headcount, however, tells only part of the story. The occupations that weakened did not all weaken in the same way. In some, the main deterioration came through fewer full-time jobs. In others, work shifted toward part time or pay weakened more than employment.
Full-Time Employment Bore the Brunt
Graphic designers and artists show some of the clearest signs of sustained weakening. Employment among graphic designers fell from 312,000 in 2023 to 294,000 in 2024 and 285,000 in 2025. Artists and related workers followed almost the same trajectory, declining from 334,000 to 318,000 to 302,000. In both occupations, employment fell in successive years.
More strikingly, almost all of the adjustment came out of full-time work. Graphic designers lost 28,000 full-time workers between 2023 and 2025, while part-time employment barely moved. Artists lost 30,000 full-time workers, again with little change in part-time work. Those who remained in full-time jobs continued to work roughly the same hours. The adjustment came through fewer full-time jobs, not shorter schedules.
Pay weakened as well. In 2023, full-time graphic designers earned 12 percent more than the national median for full-time wage and salary workers. By 2025, that premium had nearly disappeared, falling to 2 percent relative to the then national median of $62,608. Artists retained a larger advantage, but their earnings premium narrowed from 38 percent to 26 percent.
For both occupations, the signals lined up: fewer workers, fewer full-time jobs and weaker relative pay.
Full-Time Work Gave Way to Part-Time Work
Writers and musicians weakened in a different way: full-time work gave way to part-time work. Employment among writers and authors fell 14 percent between 2023 and 2025, from 271,000 to 234,000. Among writers at work, full-time employment fell by 32,000 while part-time employment rose by 7,000, pushing the full-time share down from 73 percent to 67 percent. Pay held up much better. Median earnings for full-time writers remained about 20 percent above the national median, roughly where they had been in 2023. For writers, the deterioration was concentrated in access to full-time work rather than the pay of those who retained it.
Musicians and singers show the same pattern even more starkly. Total employment barely changed between 2023 and 2025, yet full-time employment fell by 14,000 while part-time employment rose by 11,000. The share working full time dropped from 56 percent to 48 percent. Relative earnings weakened substantially as well. A stable headcount, in other words, concealed a meaningful deterioration in the structure of employment.
Pay Weakened Faster Than Employment
Photography presents a third pattern. Employment ended 2025 about 5 percent below its 2023 level. Pay moved more decisively downward. Median annual earnings for full-time photographers fell from $59,020 in 2023 to $53,144 in 2025, taking them from roughly the national median to about 15 percent below it.
Editors also ended the period weaker. Employment in 2025 was 10 percent below its 2023 level. Pay weakened too: full-time editors went from earning 29 percent more than the national median in 2023 to 15 percent more in 2025.
Across these occupations, deterioration took three forms. Graphic designers and artists lost full-time jobs and ground on pay. Writers and musicians shifted toward part-time work. Photographers saw pay deteriorate much more than employment, while editors lost ground on both.
Job losses, in other words, were only one way creative labor markets weakened.
Other Creative Labor Markets Strengthened
Not all creative occupations weakened. Some expanded substantially.
Interior designers provide the clearest example. Employment rose 21 percent between 2023 and 2025, with gains in both years. Most of the increase came through full-time work. Hours held steady, while earnings improved modestly relative to the national median.
Architects and producers and directors also ended 2025 well above their 2023 employment levels. Employment among architects rose 25 percent; among producers and directors, 29 percent. Much of that growth came in 2025, so it is too early to know whether they mark a lasting shift. But the broader indicators were strong. Among architects at work, almost all of the increase came through full-time employment. Hours were unchanged, and earnings remained roughly two-thirds above the national median. Producers and directors also added mainly full-time workers, while their earnings premium rose from 31 percent to 49 percent.
Why Generative AI Affects Creative Jobs Differently
Not every change between 2023 and 2025 can be attributed to AI. But the pattern across occupations offers clues about why the technology may matter more in some jobs than in others. A job is a bundle of tasks, and AI rarely affects all of them equally. What matters is not simply whether a technology can perform part of the work, but how central that work is to the job. For a fuller explanation of how generative AI produces creative output, see How Generative AI Works: What Creative Professionals Need to Understand.
Automating a task that takes up 10 percent or 15 percent of someone’s time may simply make the worker more productive. The consequences are greater when AI can perform a large share of the work, especially when that work is close to what the customer is actually paying for. Using AI to draft an internal memo may make a lawyer more efficient. Using it to produce the finished illustration a client would otherwise have commissioned can eliminate the need for the artist altogether.
What remains after AI performs some tasks matters just as much. Many jobs depend on judgment, coordination, synthesis, client relationships and responsibility for delivering a finished result. AI may remove some tasks while leaving those broader responsibilities intact. In other occupations, producing the text, image or other output is much closer to the economic core of the job.
Demand can complicate the picture further. If AI makes production cheaper, customers may buy more, allowing productivity and employment to rise together. But where demand is limited and AI can replace a substantial share of the labor needed to produce what customers want, fewer workers may be required.
The evidence across occupations is consistent with these distinctions. Consider graphic designers and interior designers. Both work in design, and both can use generative AI. Yet the technology reaches very different parts of their jobs. AI can substitute directly for some of the output a graphic designer is paid to produce. An interior designer still has to reconcile space, materials, budgets, contractors, client preferences and physical execution. Between 2023 and 2025, graphic-design employment fell 9 percent, while interior-design employment rose 21 percent. The contrast does not establish that AI caused either movement, but it is consistent with the idea that substitution matters more than exposure alone.
Law and accounting reinforce the point. Both were widely viewed as highly exposed to AI because much of the work involves language, documents, and analysis. Yet legal employment was broadly stable between 2023 and 2025, while employment among accountants and auditors rose 8 percent, with most of the increase in full-time work. High exposure to AI did not translate mechanically into job losses.
The important question is not whether AI can enter a job, but what it can replace once it gets there.
AI is making competent creative output cheaper and more abundant. That changes where value lies. As production becomes easier, the premium shifts toward what remains hard to automate: judgment, taste, coordination, trust and responsibility for the finished result. The emerging divide in creative work may therefore be less about who uses AI than about who still controls something scarce: a trusted relationship, distinctive judgment, responsibility for execution, or a point of view that customers value.
About Doug Woodham
Doug Woodham is an economist, former Partner at McKinsey & Company, former President of Christie’s Americas, and the author of two books about art and creative markets. At McKinsey, he helped co-found the firm’s Business Technology Office and advised companies on strategy, technology, and organizational performance. He holds a Ph.D. in economics from the University of Michigan, where his training included econometrics and the study of how innovation reshapes firms and markets. He writes about how artificial intelligence is changing the economics of creative work, drawing on his experience in economics, technology, business strategy, and creative industries.
Appendix: Defining and Measuring Creative Labor Across Occupations
This appendix explains how the creative workforce is defined, which Bureau of Labor Statistics data are used, and how employment, hours worked, and earnings are measured.
Data sources and scope
The analysis relies primarily on the Bureau of Labor Statistics’ Current Population Survey, the monthly household survey used to measure employment in the United States. Unlike the establishment survey, which collects information from employers about workers on their payrolls, the household survey includes both wage-and-salary workers and the self-employed. That makes it particularly useful for creative occupations, where self-employment is common.
The BLS classifies workers by occupation using a system of detailed occupational codes. Historically, creative occupations appeared in BLS reports mainly within the catch-all category “arts, design, entertainment, sports, and media occupations,” a grouping so broad that it obscured more than it revealed. In 2020 the BLS began publishing data for 31 subcategories within this group, providing a clearer view of the professions that make up the modern creative workforce. From these, I selected 15 occupations in which creating or shaping images, language, performances or designs is central to the work: architects; artists and related workers; fashion designers; floral designers; graphic designers; interior designers; other designers; actors; producers and directors; dancers and choreographers; music directors and composers; musicians and singers; writers and authors; editors; and photographers.
Measuring Employment and Work Intensity
Employment figures are taken from BLS Table 11, which is available on the BLS website, and represent annual averages derived from the monthly Current Population Survey. Each monthly survey asks about employment during a single reference week. People with more than one job are counted only once and are classified according to the occupation in which they worked the most hours.
Full-time and part-time employment and hours worked are drawn from Table 9b, Persons at Work and Average Hours Worked by Detailed Occupation and Class of Worker, which is available from the BLS upon request. These data differ slightly from the headline employment figures because they cover people who were actually at work during the survey reference week rather than, for example, being on vacation. Workers are classified as full time or part time according to the number of hours they usually work each week across all jobs. Those who usually work 35 hours or more are classified as full time; those who usually work fewer than 35 hours are classified as part time.
Measuring earnings
Median earnings are taken from Table A-26, Usual Weekly Earnings of Employed Full-Time Wage and Salary Workers by Detailed Occupation and Sex. Ideally, the analysis would also include comparable earnings data for the self-employed, who make up an important share of many creative occupations. Unfortunately, the BLS does not publish earnings by detailed occupation for the self-employed, making Table A-26 the only available source at this level of occupational detail. As a result, the earnings figures used here cover wage-and-salary workers only and exclude the self-employed.
Excluding the self-employed could affect measured earnings in either direction. Some independent creatives may accept lower earnings in exchange for autonomy, flexibility or creative control; others may earn more by working across multiple clients, projects or revenue streams. The BLS data do not allow those possibilities to be measured consistently across the detailed occupation codes.
To put earnings in context, each occupation’s median is compared with median earnings for all full-time wage-and-salary workers in the United States. An index above 100 means the occupation’s median exceeds the national median; an index below 100 means it falls below it. The national median rose from $58,084 in 2023 to $60,268 in 2024 and $62,608 in 2025. This provides a consistent measure of how pay in each creative occupation changed relative to the broader labor market.
