Paola Tubaro (CNRS) has published a new book chapter examining the critical role of human labor in AI development. The work appears in the book “Ce qui échappe à l’intelligence artificielle” (What Escapes Artificial Intelligence) published by Hermann.
To cite this work:
Tubaro, P. (2024). Algorithmes, inégalités, et les « humains dans la boucle ». In F. Levin & E. Ollion (Eds.), Ce qui échappe à l’intelligence artificielle (pp. 181-203). Hermann. ISBN: 979 1 0370 3844 9.
In this contribution, Tubaro challenges the common perception of AI as a purely automated technology by highlighting the essential role of “humans in the loop” – the workers who perform crucial but poorly compensated microtasks in AI development. The research makes several important contributions to our understanding of AI production and its social implications.
First, the paper examines how AI’s biases and discriminatory outcomes are deeply rooted in its production process, not just its deployment. By analyzing the global AI labor chain, Tubaro reveals how workplace inequalities and poor working conditions upstream can manifest as algorithmic biases downstream.
Second, the research breaks new ground by exposing the global scope of AI’s labor force, particularly highlighting Latin America’s role and Venezuela’s overrepresentation in this hidden workforce. Tubaro describes how Venezuela’s combination of high digital literacy and economic crisis has created a pool of highly educated workers performing essential but undervalued AI development tasks for Northern technology companies.
