The distraction of copyright regulation of AI

Those working in the creative industries have nurtured specific concerns in relation to the impact of GenAI on their work; in the legal scholarship, these have been framed mainly around copyright. That framing is understandable. Copyright infrastructures are important for the creative industries, and it is a valuable endeavour to consider how possible changes to the copyright framework to deal with the realities of GenAI will affect those industries. Focusing primarily on copyright policy is nonetheless misguided because it risks obscuring a more comprehensive description and understanding of issues and solutions.

This talk will make that broader picture visible. There are limits of copyright in relation to GenAI: the legal norms of exclusivity that underpin copyright do not neatly map on to the complexities of how data is used in the process of development of GenAI, and copyright’s territorial nature and the fact that GenAI development depends on multiple other factors aside from copyright means that it has limited regulatory influence. And even assuming a copyright licensing market for GenAI emerges, the assumption that this will stem the economic losses of reduced income for individual creators is unrealistic. Moreover, we leave unaltered existing data capture dynamics and platform dependencies.

Therefore, when it comes to GenAI regulation, should not fixate on copyright policy only, but instead put energy into exploring alternative, more encompassing solutions that might more directly tackle the complex concerns at play, and generally be attentive to those that centre the participation of creators in their design.