Work at the Frontier:
AI’s effect on freelance work
New research from OpenAI and Upwork finds shifting demand for freelance services, new combinations of skills, and larger budgets for AI-related projects.
For freelancers, changes in the market arrive one assignment at a time: a client needs a different service, requests a new skill or expands a project’s scope. Our new research with Upwork offers an early view of these changes in the AI era. Some familiar services are losing market share, while AI-related projects bring together different capabilities—and involve larger budgets and more hours of work.
The stakes are substantial. Upwork’s 2026 Future Workforce Index(opens in a new window) found that 38% of skilled U.S. knowledge workers freelance. Its previous annual study(opens in a new window) estimated that freelance knowledge work generated $1.5 trillion in U.S. earnings in 2024. Those estimates extend well beyond Upwork’s own marketplace.
Freelance work can also reveal changes sooner than economy-wide employment statistics. Contracts are often short, and clients can adjust their requirements with each posting. To get to the root of these changes, we partnered with Upwork, an online marketplace connecting clients and freelancers, to study those requests alongside the activities people undertake directly with ChatGPT.
The mix of work differed across the two platforms. Writing accounted for a much larger share of ChatGPT use, while video and animation accounted for larger shares of requests for freelancers in Upwork postings.
Below are the ranges of tasks between the two platforms.
These activities represent only the largest changes in Upwork shares over the studied period. ChatGPT shares cover Jan 2025–Jun 2026; Upwork shares are H1 2026. Changes in Upwork share are H1 2025 to H1 2026, in percentage points.
Many activities that were relatively common in ChatGPT lost share on Upwork, while less common activities gained share. Of the 63 work categories analyzed, 43 followed this pattern.
The remaining 20 work categories showed the reverse. ChatGPT use alone does not explain changes in freelance demand.
The comparison also captures how people currently use ChatGPT, not the full range of tasks AI can support—a gap OpenAI calls the capability overhang.
Each point represents one of 63 matched work subcategories in the Upwork taxonomy. The horizontal axis is the natural log of the category’s share of ChatGPT use in H2 2025 divided by its share of Upwork job posts in H1 2023. The vertical axis shows the percentage change in the category’s share of Upwork postings from H1 2023 to H2 2025, on a logarithmic scale; 0% indicates unchanged marketplace share. The vertical divider at AI intensity 0 marks equal category shares in ChatGPT and baseline Upwork postings.
By late 2025, clients were requesting the use of AI in many more, different kinds of work. Clients were seeking freelancers to apply AI in tasks including chatbot development, video production, marketing automation, and application integration.
Within individual projects, AI capabilities were often requested alongside other work. AI-related postings listed about as many tasks as other postings but more often combined tasks that had rarely appeared together in earlier jobs.
Some clients asked for AI-generated video, editing, and motion graphics in a single posting. The freelancer would bring those skills together to deliver a finished video.
AI-generated video
Video editing
Motion graphics
Between early 2023 and late 2025, the mix of tasks in AI-related postings became 45% more diverse—the equivalent of an increase from about 15 equally common tasks to about 21.
- Early 2023
- 14.6
- Late 2025
- 21.2
Although posted hourly rates were similar for AI-related and other work, AI-related projects had larger fixed-price budgets and involved more hours of work. This pattern suggests that higher budgets may partly reflect a broader scope of work.
AI’s effects on work can be difficult to detect in employment statistics. As we have noted in previous Work at the Frontier papers, the effects may appear first in the work people do.
Because freelance projects turn over quickly, postings are continually refreshed as clients’ needs change. Work within firms can evolve without appearing in a new job posting. Combined with ChatGPT data, these postings offer an early view of which activities people undertake directly with AI, which they hire freelancers to perform, and how AI-related projects differ from other work.
The findings do not point to a simple substitution story. Some activities common in ChatGPT use lost share on Upwork, but this pattern was not uniform. AI-related work also spanned a wider mix of activities, combined tasks in less familiar ways, and carried larger fixed-price budgets and more hours.
Together, these patterns suggest that AI may be shifting both what clients seek from freelancers and how they assemble projects.
The OpenAI Economic Research Team explores how AI is changing work in real time with the Work at the Frontier series.
