Expert Comment: How is generative AI transforming the labour market?

Dr Fabian Braesemann. Credit: University of Oxford.

To find out whether the freelance job market has changed as a result of AI and how it may be reshaped in future, I worked with an international team of economists, mathematicians, and data scientists from the University of Copenhagen, the University of Oxford, UCL, the Complexity Science Hub from Vienna, ETLA from Helsinki, the University of Cambridge, and the Einstein Center Digital Future from Berlin.

The result of this two-year research project is a new study, published in the Journal of Economic Behavior & Organization, outlining the labour market effects of generative AI. We analysed millions of data points from an online freelancing platform around the release of ChatGPT in November 2022.

Our data shows that the labour market effects of generative AI are nuanced: for some jobs, we find reduced demand, while for others, demand went up. The job types that are most affected are those where generative AI significantly impacts work routines, that is, everything from writing, accounting and programming to project management and graphic design.

Whether an occupation is at risk of possibly being replaced hinges on its substitutability or complementarity with AI. For a few job types, generative AI has become so powerful that it can effectively replace human labour; yet for many more occupations, generative AI can help employees be more productive. 

In January 2025 the stock market reacted heavily to the release of China’s generative AI tool ‘DeepSeek’, with shares in the American chip manufacturer Nvidia and other highly valued companies associated with the recent AI hype dropped dramatically. These rapid stock market movements show the market for generative AI technologies is highly dynamic. If generative AI tools continue to spread as rapidly as they do now, we might soon see large companies develop their own AI tools specifically tailored to the needs of their respective industries. The adoption of AI tools by companies will have a long-lasting impact on the job market.

Dr Fabian Braesemann, Departmental Research Lecturer in AI & Work at the Oxford Internet Institute

For example, instead of spending a long time trying to solve a tricky programming task by googling or consulting online Q&A websites, a software developer can now ask ChatGPT or DeepSeek for help. The code can simply be uploaded to the AI tool and the machine will improve the code to solve the programming task quickly. It is these time savings that can significantly increase the productivity of employees in many different industries.

But what does this development mean for the labour market as a whole? 

First, ChatGPT, DeepSeek, and other AI tools are not the big job killers that some have feared. Instead, generative AI can increase productivity if integrated smartly into business operations. The increased demand for chatbot development and machine learning jobs we see in the data points to the potential for tapping into new markets and product developments by utilising generative AI.

Organisations that successfully integrate AI applications productively into their business practices will gain a significant competitive advantage. Against this backdrop, we—the users of AI tools—can be happy about the new competition between American and Chinese AI companies: the competitive pressure will lead to more innovation and lower prices for AI products and applications.

DeepSeek is also breaking new ground by making its AI tool available under an ‘open source’ licence, which should significantly increase the spread of AI applications in the economy. In contrast to other AI tools, users can customise DeepSeek as they wish and access the software's source code. It is expected that innovative companies will not miss this opportunity and will now try harder to generate new AI applications that could either make existing processes more efficient or open up entirely new markets.

The current dynamic market movements show that the generative AI hype is far from over. With every new model and software release, companies worldwide are learning more about how to set up effective AI systems. If generative AI tools continue to spread as rapidly as they do now, we might soon see large companies develop their own AI tools specifically tailored to the needs of their respective industries.

To harness the productivity and innovative potentials that generative AI promises, all economic stakeholders need to work together: firms need to establish practices on using generative AI effectively in the workplace. The workforce must be familiarised with the productive use of AI, and business processes must be adapted and modernised. Employees should embrace learning new skills that help them strengthen complementarities with state-of-the-art tools, and educational institutions should support that skill transformation. Governments should create regulatory frameworks that allow companies, employees, and the public sector to use generative AI productively.