From the Hollywood writers’ strike of 2023 to the recent announcement by Disney that it plans to lean heavily into AI, there is growing concern that this new technology will negatively impact the market for human creative work. A new study by Professor Kristofer Erickson, published this week in Creative Industries Journal, sheds light on this complex issue with empirical evidence from firms that are working with AI.
The article reports on a multiple case study analysis of small creative firms that commercialised products made with AI input. Interviews with key decision makers and production staff reveal how AI was perceived within firms and how creative work was impacted by introduction of the technology. The study covers a diverse set of creative industries: photography, music, video games, choreography, publishing and architecture.
The main finding was that AI did not directly replace human workers, but required flexibility and new skills, including computational expertise. Significant additional labour was involved in locating, cleaning and preparing input data for AI training. Overall, the study found that AI was being used as an ideation tool, allowing firms to quickly explore multiple creative possibilities before committing to one. Firms’ commercialisation strategies often followed a platform model, and frequently involved a high proportion of audience/user input.
Erickson argues that understanding the effects of AI is best informed by looking at continued processes of digitalisation which have already been transforming creative work. The end result is therefore complex. AI is displacing work by humans in certain areas, while opening up new creative expressive possibilities, all within an environment of significant uncertainty and risk:
Product teams were small, flexible, distributed and subject to considerable risk, all characteristics identified in previous studies of digital creative SMEs. Case study firms embraced AI because it offered new possibilities for creative expression, but this brought new complexities and knowledge requirements, often requiring collaboration across sectors or beyond the boundaries of the firm. […] Firms used AI as a means of exploring the interface between computer and human creativity, and to reflect back like a mirror on the styles, conventions and genres of previous creative practice underlying their artforms. (p. 17)
Copyright and legal scholars will be interested in what the case studies had to say about licensing of input data. This caused frequent impediments and uncertainties, preventing firms from fully commercialising downstream outputs. Often, firms felt as though they were operating in a regulatory “grey area”, absent legal clarity about their own liability or the legal status of training materials.
To read the full open access version of the article, please click below.
Stay tuned for more of our team’s publications and research updates by subscribing to our newsletter.
