Working for AI

This chapter examines the role that platform labor plays in the development of contemporary AI systems. While such systems are often touted as magical innovations, they are typically propped up by vast amounts of human laborers, who clean the data, manually label key features, and moderate toxic content, among other tasks. Proponents claim these tasks offer flexibility and pay; critics counter that this work is exploitative and precarious, taking advantage of the already marginalized. This chapter surfaces this often-invisible labor, highlighting several key issues around its poor or nonexistent remuneration, exploitative mechanisms, negative impact on well-being, and extractive colonial logics. The chapter suggests several interventions, from concrete policy to corporate responsibility, that might lead to improvements. As AI technologies proliferate into many domains, the hidden labor required to develop them—and the negative impacts this has on lives and livelihoods—becomes an increasingly urgent issue.

Read the book chapter (open access)