Monsters of Our Own Creation: AI, Occupational Cannibalization, and the Future of Work

While ethnographically studying a startup developing an artificial intelligence (AI) technology, I puzzled over the phenomenon of “occupational cannibalization”: the fact that occupational members who ran the startup – namely, music composers – ended up developing an AI intended to carve away at their community’s work of music composition. Initially, those at the startup developed the AI to keep this work within the hands of members of the occupation, against its complete takeover by machines. This said, the AI started to be used in ways that did not allow for this imagined future. Instead, the people who started to use it – namely, video content producers – were able to produce music for their videos on their own through the AI, instead of having to hire the music composers they traditionally relied on. In other words, and contrary with the initial intentions of those developing the technology, the AI ended up being used to completely replace some of the occupation’s members. However, and as argued by composers at the startup and beyond it, these emerging patterns of use were consistent with hierarchies of worth within the occupation: the work being automated away, given that it was not seen as being “soulful” in the ways that composers cherished, was not seen as worthy enough to be saved by the occupation’s members. The composers both at the startup and beyond thus were willing to relinquish this work, as well as those who did it, across the community’s boundaries to the advance of machines.