The Future of Creative Work
How AI Is redistributing creative work
We studied two waves of US job postings, surveyed 1,433 working creatives and 3,300 people training to enter the field, and reviewed eighteen months of qualitative research. The sources illuminated the same general pattern: AI is reshaping creative work and the people who do it.
Stay in the loop
Subscribe | Subscribe - AI thought leadership Email us | Email us - AI thought leadership
Stay in the loop
We publish new AI research as it's ready; no marketing, no hype. Subscribe to get it in your inbox, or reach out anytime with a collaboration pitch, idea, or feedback. We read it all.
Subscribe | Subscribe - AI thought leadership Email us | Email us - AI thought leadership
JUMP TO SECTION
AI is making first-pass work much easier. Judgment is filling the space
Clients are expecting faster work but setting limits on how AI can be used
AI skills are valued more by larger companies than smaller ones
Creative professionals worry AI may negatively impact their career. Entrants see it as the way in
Abstract
Over the past year, we studied how creative work and the people who do it are changing. Across two waves of US job-postings data separated by six months, a survey of 1,433 creative pros and creators, an eleven-country study of 3,300 people training to enter creative work, and qualitative surveys with working creatives, we have found evidence of redistribution rather than replacement. Creative professional postings rose in our six-month comparison, and AI skills became more explicit in job requirements. Working creatives have incorporated AI into specific parts of their workflow, while drawing sharp boundaries around other parts that give creative output its sense of humanity and ownership.
The sharpest divide in the data was not between disciplines or roles, but between established creatives and those entering the profession. In some areas, working creatives are setting boundaries that exclude AI from parts of their workflow, while those entering the profession are using AI more broadly, and as part of their path into creative fields.
As AI continues to evolve, and creatives increase their comfort level with it, our research suggests the emphasis may shift. The question may become less about how effective AI is at producing creative output in isolation, and more about how it supports creatives in directing, evaluating, and accounting for their work.
This article is the first in a new series of research reports about creative work in the age of AI.
Key Findings
- Creative hiring is up. US creative professional job postings rose 8% between September 2025 and April 2026, Desire for AI skills is heaviest among mid-market and large enterprise employers.
- Working creatives are optimistic about AI but sensitive to how it’s used. Eight in ten US creative professionals (84.8%) feel positively about AI, particularly for brainstorming.
- Clients are expecting faster work for some deliverables. Practitioners describe pitch deadlines shrinking from weeks to days, and a growing tier of acceptable-but-unremarkable AI work overtaking smaller projects like social media posts.
- More people are seeking creative jobs, boosted by AI. About twelve million Americans are training to move into creative work. Three of four of them (77%) rate AI as essential to that path.
- Creative professionals are using AI to check the originality of their work. In interviews, some creative professionals describe deliberately checking their own ideas against AI output and starting over if AI generates the same idea.
Introduction
There are two primary narratives about AI and creative work. In one, creative professionals are being displaced as generative tools absorb more of the production process. In the other, creative capacity is expanding as more people gain access to professional-grade tools. Anecdotal evidence indicates there may be some truth to each story. But neither narrative tells us much about what is happening inside briefs, how workflows and hiring requirements are shifting, or the daily choices working creatives are making.
We set out to study these deeper questions across four sources of evidence collected by Adobe. We consulted two waves of US job-postings data, a 1,433-person survey of creatives in the US and UK, an eleven-country study of 3,300 people actively training to enter creative work, and a body of qualitative studies conducted by Adobe Design Research & Strategy (ADRS) over the last 18 months.
The answers were at times surprising. In the evidence we examined, demand for dedicated creative roles rose. AI requirements became more explicit. And creative labor appeared to be moving away from producing first-pass variation and toward selection, direction, client negotiation, and accountability for how work was made.
What we mean by creative work
Throughout this report, we distinguish between three groups.
- Creative professionals are people whose primary paid work is in design, video, photography, illustration, motion, or related creative disciplines.
- Creators are hobbyists and other people producing creative work, often publicly or commercially, but not necessarily as their primary occupation.
- Creative entrants are people actively training to move into creative professional work within the next 24 months.
The boundaries between these groups are blurry, and AI is making them more so. We separate them here because the evidence suggests each group is experiencing AI differently. Incumbents are integrating AI into the craft they already practice. Entrants are using AI to access a craft they are trying to enter.
AI is making first-pass work much easier. Judgment is filling the space.
— Graphic designer
Looking at task-level data from creatives about how their roles are changing we found that AI is used predominantly in two parts of the creative workflow: the front, where it expands creative possibility, and the back, where it absorbs the post-production work that lives between a finished idea and a delivered asset.
Among US creative pros, 84.8% report positive sentiment about AI's effect on brainstorming and ideation, the highest sentiment score on the survey. In the same survey, 81.7% rated brainstorming and ideation as having high AI enablement, meaning AI has risen to the requisite skill level for the task. The pattern is repeated in creators: 91.6% of US creators rated brainstorming as enjoyable, 84.7% rated it as enabled by AI. Brainstorming is the only task in the survey that simultaneously scored high on creativity, high on enjoyment, high on AI use, and high on positive sentiment about AI's role.
AI is used in the later stages of workflows most clearly for photo and video finishing. Among US photo professionals, 57.6% report substantial use of generative AI to remove distracting background elements, 50.0% on compositing multiple images, and 45.1% on color and tone retouching. Among US video creators, 48.6% report substantial generative AI use on visual effects.
As the beginning and end stages have become optimized by AI, the middle stage tasks have expanded. These tasks include selection, direction, client communication, and accountability for what was made. The same redistribution pattern shows up in qualitative interviews across multiple disciplines. For instance, a graphic designer we interviewed described the redistribution as a question of where her value sits.
Regarding productivity: we observed that AI is being used for parts of creative work that practitioners describe as time-consuming but not where their craft is most expressed. The parts AI typically does not absorb are the parts practitioners describe as where their value lives.
What did not change may be as informative as what did. Tasks involving physical capture and collaboration have lower AI adoption. Among US video professionals, only 8.3% report substantial AI use in planning the setup of camera and lighting equipment; 13.3% report AI use in filming footage or capturing shots. Among US photo professionals, 8.1% use AI in setting up camera and lighting equipment. The same pattern is visible with tasks that involve human collaboration. Among US creative professionals across all disciplines, less than one in four (24.1%) use generative AI in reviews with collaborators. The boundary is consistent across diverse roles and across the US and UK samples.
Clients are expecting faster work but setting limits on how AI can be used
— Art director
Most commentary suggests widespread reduction in budgets alongside lower quality thresholds, but our qualitative research describes something different and more specific. Clients are simultaneously tightening restrictions on where AI can be used in their commercial output and compressing the time available to deliver everything else.
The clearest enterprise pattern is restriction on where AI can be used. For example, one art director we spoke to described the rules her team works under, noting that AI is acceptable to render ideas and mock-ups, but not allowed in final work.
The pattern is consistent with the practitioner-side data. Three-quarters of creative professionals in the US and UK agree that AI models should not be trained on artists' work without consent: 70.5% in the US and 74.9% in the UK. Concerns about copyright and IP are also significant: 56.7% of UK creative professionals report concern about copyright or trademark infringement when using AI outputs. Our research showed some evidence of this influencing creative behavior. For example, one art director shared that her company’s policy allowed AI to be used for final work only if the models used were trained ethically and did not to infringe on copyrights.
A second pattern suggests intermediate deadlines are compressing. A film director and producer we interviewed described the time pressure on commercial pitches as the part of the work AI is most needed to relieve:
— Film director & producer
A third pattern is also emerging. Working creatives report that client expectations on lower-stakes deliverables were eroding in some places. The clearest example named in our surveys was social media: several respondents reported doing fewer social media deliverables because some clients are now generating that work themselves. This was described as a growing tolerance for 'AI mid' output—AI work that is acceptable rather than excellent.
The work clients are paying for was also clear in our analysis. Value remains steady for high stakes outputs, such as commercial ads and customer-facing creative. Respondents also reported consistent demand for work AI is least able to do on its own, including final commercial output with verifiable provenance, taste at the pitch and selection stage, and quality work delivered against a compressed clock. (Adobe Research plans to study these trends in greater detail and report on them in future reports).
AI skills are valued more by larger companies than smaller ones
— Brand designer
Two waves of US job postings for creative professionals showed two intriguing patterns.. The first is that the value of creative work is increasing. The second is that AI requirements are concentrating in larger employers.
Between our September 2025 and April 2026 datasets, US creative professional job postings rose from roughly 10,500 to 11,300, an 8% gain over six months. The share of postings that explicitly listed AI skills rose from 10% to 15%.
Over the same period creative work appears to have shifted back toward dedicated creative professional roles rather than dispersing across general business roles. Two years ago, the typical marketing or business-operations role increasingly listed Photoshop, Premiere, or Figma in its requirements, and the generalization of that trend produced confident forecasts about a future in which everyone will be a designer. Our April 2026 wave shows the movement reversing.
There was also an overall 5-point rise in AI as an explicit requirement in creative job postings. This requirement is found across multiple different creative professional job categories: Graphic Designer accounts for 11% of AI-requiring postings, Video Editor 10%, UX/UI Designer 8%, Animator, Photographer, and Motion Graphics Designer 6% each, and Art Director 5%. These are not new roles. They are existing roles with a new line in the job-requirements section.
The second pattern is in the size of companies asking for AI fluency. Among job postings from mid-market and enterprise companies, 18% explicitly required AI skills as of April 2026. Among postings from solo and small businesses, only 8% did. Mid-market and enterprise share of creative professional hiring also rose from 64% to 69% of postings over the six-month window, meaning AI requirements are concentrating in the segment of the labor market doing more of the hiring overall.
The supply side of this market is registering the shift. Respondents to our surveys described AI fluency as a new credential that is becoming crucial for creative entrants. Whether AI skills are expected in the companies where they may eventually work, these early-career creative entrants expect that having AI fluency will make them more competitive candidates.
Creative professionals worry AI may negatively impact their career. Entrants see it as the way in.
— Creator, entrant
Working creatives report that AI is changing how they work, creating opportunities for iteration and improvement. People training to become working creatives see it as a path into the field.
About twelve million American adults, 4.4% of the online-connected US population, are actively training to become creative professionals. That is about three times the size of the current US creative professional and creative-major-student population combined. According to that study, 93% are career switchers transitioning from non-creative jobs. The most common current occupations of people trying to shift into creative roles are engineering and architecture (11%), accounting and finance (10%), business management (10%), IT and software (9%), and administrative or office support (9%). More than half (56%) have no formal creative education. Three in four (77%) rate generative AI as “extremely” or “very” important to their future creative career.
The incumbent view is more nuanced. In one of our surveys, working creatives were 85.6% net-positive about AI, and at the same time, 46.2% were likely to call it a threat to their profession.[2-7] We do not read these findings as contradictory. They suggest that AI can be useful inside a workflow and still feel threatening to the structure of a profession. UK creative pros report a higher level of threat perception (51.2%) than US creative pros (46.2%).
In our qualitative research, much of the generational difference appeared in the form of differing assumptions about what AI fluency is. To a 17-year-old creator we interviewed, AI tools were not a credential to acquire. They were the working environment.
For an incumbent re-entering the job market, the same fluency is something to retrofit, often against the resistance of peers:
— Brand designer
Incumbents and entrants are reading the same technology from different positions. Incumbents see AI as something to integrate into a craft they already practice. Entrants see AI as the price of admission to a craft they are trying to enter. Many of those entrants are arriving from technical or business backgrounds, which may give them an initial fluency with AI-mediated workflows that existing creative professionals had to retrofit.
The labor-market story over the next 24 months may reflect the changing entry path into creative work. A risk for incumbents may be being outcompeted by entrants whose AI fluency is native rather than retrofitted. A risk for entrants may be mistaking AI fluency for the craft itself.
Working creatives are drawing red lines they won’t let AI cross
— Creative director
The rejection of AI for some uses is forcing more human creativity, not less.
A consistent body of practitioner research shows that working creatives are adopting AI selectively, but they are also actively pushing back on it in specific areas and developing techniques to keep AI out of the parts of their work that feel most their own.
Our surveys revealed places in the workflow where practitioners are not actively using AI even when it is technically capable. This includes tasks that could disrupt authenticity, artistic integrity, ownership, or emotionally impactful design. Empathizing with clients was also an area where creative professionals were not willing to take shortcuts. Overall, creatives were reluctant to fully automate steps of the workflow at this time, preferring to keep humans in the loop.
Similar logic appears in other data. In additional research, creative professionals reported that prompting an AI model to make work does not produce the same sense of pride of authorship that drawing, painting, or photographing does. A graphic designer we interviewed put the feeling plainly:
— Graphic Designer
— Creative director
The hiring data echoes this from a different angle. As of April 2026, 85% of US creative professional postings still do not explicitly require AI skills. Five out of six creative jobs being posted today are still being filled, in formal terms, without a requirement for AI fluency. The market has not converged on AI as a baseline credential; it is concentrating AI demand in specific segments and leaving the rest largely untouched.
What working creatives appear to be doing is harder to summarize than as either “embracing AI” or "avoiding AI." They are integrating AI in specific zones, restricting it in others, and developing their own techniques to make sure the work that bears their name is still recognizably theirs. Evidence suggests the balance may be selective integration paired with deliberate resistance.
What this may mean
The findings in this report are based on limited data and warrant additional research. But our evidence points in a broad general direction. In the six-month hiring comparison, creative professional postings rose, dedicated creative roles gained share, and AI became more explicit in job requirements. In the survey data, working creatives reported a clear pattern of where AI is actively used and where it is not. In our qualitative research, creative professionals and entrants described the same boundary in their own words, often paired with explicit techniques for keeping AI out of the parts of their work they consider most their own.
We read this convergence as redistribution of creative work, not a replacement of creative workers. AI appears to be making some parts of the creative process faster and easier, particularly brainstorming, first-pass variation and late-stage finishing. Our research suggests that the in-between work is moving into selection, direction, client negotiation, accountability for how work was made, and the active practice of guarding originality.
These findings suggest several implications for the people building, using, and governing creative AI systems. For teams managing creative work, the most visible change may not be shorter projects but a redistribution of how time is allocated in creative projects. For builders of creative AI tools, the findings suggest that selection, comparison, rights review, and collaboration are likely to become as important as generation. The most valuable skills are likely to remain creative judgment and vision.
The entrant data complicates the picture in a productive way. Working creatives are broadly positive about AI while also likely to see it as a threat to the profession. People training to enter creative work are more likely to treat AI as part of the path into the field. The next phase may appear like a traditional labor transition, less about immediate displacement than about a generational handover in which the entering cohort will be more AI-fluent.
It is important to note that this report does not predict long-term creative employment. It does not estimate aggregate productivity gains. It does not measure the share of practitioners building their own multi-tool AI workflows. It does not offer judgement about the quality of AI-augmented creative work when compared to non-AI work. We intend to further study those questions with credible data and qualitative research from creative professionals and entrants in creative fields.
The narrower point the evidence supports is that more creative options are unlikely to be the scarce input in creative work. Instead, the scarce input appears to be the judgment to decide which options are worth pursuing, and the deliberate practice of keeping enough humanity and human creativity in creative work.
Methodology
Job postings analysis. We reviewed two waves of US data. The September 2025 wave drew on roughly 28,000 creative-app postings from Indeed, Behance, and Dribbble via Apify, and the April 2026 wave drew on approximately 25,000 postings collected by Revelio Labs with roughly 11,300 classified under creative professional categories. The methodology shift between waves was forced by a policy change at Indeed. AI workflow skills were operationalized by tagging postings whose required skills text matched a controlled list of generative AI-related terms. We acknowledge postings are not the same as hires, and that skill tags in job postings may not accurately represent a role’s complete scope.
Survey of working creatives. We consulted a brand-blind survey conducted by Adobe Design Research & Strategy and fielded between April 27 to 30, 2026. Total n=1,433 across US and UK creative professionals and creators: US creative pros 403, UK creative pros 203, US creators 485, UK creators 342. In the survey, task-level questions captured generative AI use, sentiment, and AI enablement across discipline-specific task lists. Attitude questions captured overall AI sentiment, perceived efficiency and quality effects, consent and IP attitudes, and threat perception. This was a cross-sectional snapshot, not a long-term trend.
Study of creative job seekers. We reviewed a study of job seekers conducted by ADRS in October 2025 and published in November 2025. The 8-minute online survey was conducted across eleven countries, n=300 per country (total n=3,300), and screened for active intent to transition into a creative profession within 24 months. The twelve-million figure quoted in this report is a US population projection based on the screen incidence rate.
Qualitative practitioner research. Testimonials and quotes draw on multiple ADRS qualitative studies conducted between 2025 and 2026. Quotes throughout this report are taken directly from these studies and reproduced with their original attribution. Names have been removed to protect the anonymity of respondents.
What this report cannot tell you. This report draws on a six-month delta in hiring data, not a multi-year trend. We do not have time-allocation panel data on how project time is shifting within creative workflows. The redistribution finding is a qualitative pattern observation supported by cross-sectional task-level survey data, not direct time-use measurement. These and related questions will require additional research.
Acknowledgments
This report was produced by Adobe Research Teams, with strategy and production by Mike Downey, Hala Anwar, Paul Slater, and Daniel Stone. Research and analysis contributions were collected and produced by researchers with Adobe Design Research & Strategy and by Adobe Brand Strategy & Insights, including Zihan Miao, Kimberly Welchons, Nick Brown, Bazile Lanneau, Jessica Outlaw, Victoria Hollis, Wilson Chan, and Jenna Melnyk. We thank the 1,433 respondents to the Evolving Roles survey, the 3,300 respondents to the Job-Seeker study, and the hundreds of working creatives and creators who have participated in ADRS qualitative research from which the quotes in this report are drawn.
References and notes
US creative professional job posting analyses, conducted by Adobe Brand Strategy & Insights in two waves: September 2025 (28,000 postings from Indeed, Behance, Dribbble collected by Apify) and April 2026 (25,000 postings from Revelio Labs, with 11,300 classified as creative professional).
Adobe studies and surveys with creative professionals and creative entrants, conducted in 2025 and 2026.