The Mythical Speech of Artificial Intelligence

The media narrative surrounding the present and future capabilities of AI often misattributes intelligence to computational systems. This text argues that this narrative is grounded on a mythical speech that assumes intelligence stems from autonomous agency—leading to the belief that if computational systems solve problems autonomously, they must therefore be intelligent like humans. Drawing on ethnographic research conducted with a deep learning engineer, the chapter shows that while we should not deny the agency of AI or its current role in human cognition, this does not mean it is an autonomous agent—much less one that possesses human-like intelligence. On the contrary, this ethnography will reveal that the intelligence computationally simulated by deep learning systems is not the result of autonomous agency but of social agency. Lastly, the argument draws on an experimental ethnography which aimed to incorporate social agency into the design of human–AI interfaces as a way to resist the AI mythical speech. The goal is to propose an interventionist and forward-looking anthropology with AI—one that enables us to speculate about alternative ways of imagining the future of this technology, beyond the dominant narratives that accompany its development.