CREATe is happy to present the eight entry in our series of working papers released in 2024: ‘AI: A cure for Baumol’s disease?’ by Gillian Doyle (Professor of Media Economics, CCPR, University of Glasgow) and Sabine Baumann (Professor for Digital Business at Berlin School of Economics and Law, Scientific Director at OFFIS Institute for Information Technology).
Gillian and Sabine write:
The production of outputs in cultural industries, including film and television, is said to suffer from Baumol’s disease in that, because of its emphasis on creative labour elements which cannot readily be mechanised, replicated by computers nor streamlined, it is prone to above-average inflation. But is this still the case? Recent developments in generative Artificial Intelligence (AI), with its capacity to assist in generating text, images and sound, have produced umpteen opportunities to automate aspects of creative work across media and other cultural industries. What does this imply for Baumol’s cost disease?
Described as ‘the jewel in the crown of cultural economics’ (Blaug, 2001: 131), Baumol’s cost disease refers to the ‘inevitable’ escalation of real costs that occurs in labour-intensive industries such as the arts (Towse, 1997) where, because of the emphasis on certain specific sorts of creative labour, ‘flexibility does not exist’ to substitute capital for this input (Towse, 2020: 416). Processes of production of cultural outputs, including the making of film and television content, are susceptible to the disease ‘because creativity is inherently labour-intensive and because labour costs tend to rise more quickly than others, [which means that] costs across these sectors will tend to rise at a faster rate than inflation’ (Doyle, 2013: 101).
Click on image to download.
However, recent developments in generative AI have established unprecedented opportunities for mechanisation of a range of aspects of creative work. Clearly, the capacity of generative AI to assist in generation of text, images brings with it significant potential to transform and disrupt processes of video, audio and text creation. To what extent might AI, by potentially reducing costs and raising productivity in inflation-prone media content creation activities, challenge conventional theory and effectively counteract Baumol’s disease?
While some media executives highlight concerns about how developments in AI might encourage copyright infringement by, for example, text chatbots and image generators usage of content (Criddle et al., 2023), others acknowledge the positive potential and ‘a massive opportunity to optimise creative work’ (Read, cited in Pitel and Storbeck, 2023). The embrace of AI as a tool to support creative work, although controversial, evidently offers significant opportunities to automate and underpin production of an array of creative outputs, from advertising campaigns and animations to songs and storylines for the printed page and digital screen.
Drawing on analysis of secondary source literature, particularly in the realms of media and cultural economics and media business studies, and on reports and specialist press coverage of the development of AI and of media and creative industries, our essay assesses the ways that AI is transforming processes of creativity and co-creation within media and creative sectors. We identify how use of AI to automate some of the time-consuming and repetitive tasks that are integral to content creation and production, from research and pattern identification to creation of mock-ups and tailoring, is increasingly prevalent. Our analysis highlights how career paths are being reshaped, which roles are subject to diminution, how creative processes are being enhanced and areas where new skill sets are required – see the table below.
Role of AI in different Media Industries and Effects on Jobs. Source: Doyle and Baumann, 2024
As is typical of any efficiency-improving technology, AI is cutting down on the need for human labour to perform certain tasks involved in media production. Integration of AI tools can allow journalists, story developers and other content-creators to offload some of the more time-consuming functions involved in their work, thus freeing up more time for them to engage in editorial creativity, to the benefit of both productivity and content quality. So media organisations evidently have much to gain from integrating AI tools judiciously into processes of creating and supplying content.
Does this mean that at last Baumol’s cost disease has been cured? Alas not.
Certainly, the arrival of AI challenges banishes any notion that all forms of creative labour inputs are irreplaceable. But we would argue that, even though the use of AI tools has increased significantly amongst media organisations and can yield a myriad of efficiencies and cost-savings, the core creative work involved in content creation remains heavily dependent on human creative labour. Although AI is a powerful tool, its risks and limitations are such that, inevitably, humans need to remain ‘in the loop’ (Zysman and Nitzberg, 2024). Echoing the findings of earlier research which suggests that human-AI collaboration is what drives improvements in productivity (Nah et al., 2023; Sowa, Przegalinska and Ciechanowski, 2021), our analysis of how AI tools are used in media content creation indicates that, rather than substituting for human ingenuity, they typically play a support role. For example, where AI is used in news production it generally requires extensive human intervention to safeguard and ensure the coherence and accuracy of ensuing news content. Likewise, where AI is used to help suggest storylines, plot twists and dialogues for example for television, it nonetheless requires human input and editing by creative scriptwriters and other media professionals to ensure the quality and coherence of the ensuing content. Human creativity remains at the core of content production. So creativity within media content-making industries remains inherently and obstinately labour-intensive. As a consequence, Baumol’s proposition (Baumol and Bowen, 1966) that these industries are potentially susceptible to higher than average cost inflation still holds true.
Be that as it may, the potential for generative AI to further transform media creation processes and media businesses over coming years should not be under-estimated. As AI is increasingly adopted, key challenges for media businesses include how to protect their intellectual property assets, develop specialist AI skills, balance machine learning with human creativity, and use AI ethically. The agenda for policy-makers is no less daunting, bearing in mind AI-related concerns about misinformation and transparency, copyright and licencing, potential market dominance and environmental sustainability. So there’s no doubt that the research agenda in this area will continue to evolve and flourish over coming years.
AI: A cure for Baumol’s disease?
Gillian Doyle and Sabine Baumann
CREATe Working Paper 2024/08
Abstract
The production of outputs in cultural industries, including film and television etc, is said to suffer from Baumol’s disease in that, because of its emphasis on creative labour elements which cannot readily be mechanised, replicated by computers nor streamlined, it is prone to above-average inflation. However, recent developments in generative AI, with its capacity to assist in generating text, images and sound, have established unprecedented opportunities to automate and support aspects of creative work across media industries. What does this imply for Baumol’s cost disease? This paper examines recent developments in AI, and analyses to what extent, by potentially reducing costs and raising productivity in inflation-prone media content creation activities, these technologies challenge conventional theory and effectively counteract Baumol’s disease.
Full paper can be downloaded here.